Qualcomm CEO Cristiano Amon Envisions AI Agents Replacing Mobile Apps and Smart Glasses Making Current Smartphones Obsolete Through Persistent Human Machine Interfaces
Qualcomm CEO Cristiano Amon has predicted the future of how people interact with digital services, envisioning that artificial intelligence agents will gradually substitute the concept of apps. Speaking on the CNBC Tech Download podcast, he foresaw a time when the digital assistants would perform complex tasks on the user's behalf, eliminating the need to open individual programs. Apps will likely not disappear entirely, but they will evolve and operate behind the scenes, controlled by the agents that form the human machine interface.
The emergence of autonomous agents will fundamentally alter how we view mobile software. Amon used the example of a banking transaction in which the agent would provide specific details from your account with a single voice query, rather than requiring users to navigate menus and pages within an app to find the same information. For Qualcomm, those agents "are going to be the new app"
in a new era of consumer devices.
That impending shift in how we use software is driving a new paradigm of devices, and Qualcomm is gearing up for a wide range of them, as the company has already begun working on more than 40 designs for AI devices. "We believe that smart glasses will grow and will be one of the leading categories and will scale to hundreds of millions,"
Amon reported, and that smart glasses will "make the smartphone obsolete in that form factor"
. In the past year, smartphone shipments globally totaled 1,260,000,000, a three percent growth from the previous year's shipments.
But other devices beyond augmented reality eyewear are also in development, including the chipmaker's processors for smart rings, pins, watches, and even earbuds with integrated cameras, designed around the concept of continuous human observation by hardware worn by the user, "a device that you wear, that is always with you, that can see the world around you, so you can have context and the agent will have context so that you can ask an agent something."
The agent centric hardware transition has also sparked the attention of other developers who will be bringing new hardware to the consumer electronics marketplace. Big names in AI are reportedly already partnering and buying out hardware makers to gain direct consumer access to them this will likely include startups that are currently sponsored by major AI companies, aiming to produce their own hardware. "All of the devices that we wear,"
according to Amon, "become an endpoint for the agents, and those AI companies understand they need to win those endpoints from the agents."
User data is what will be drawing software developers to hardware manufacturing; streams of real world data generated by sensors and cameras on wearables so that these developers can train hyper specific, contextually relevant AI models for the users they serve will have a vastly greater impact on model training data sets than the data used today to train static language models.
This persistent and continuous stream of computationally demanding tasks would demand an entire redesign of how semiconductor processors function and perform; current mobile processors are not designed for the constant background processing and low power consumption required for always on AI observation and would therefore require a full revamping of Qualcomm's processor line. "There's nothing on our roadmap today,"
confirmed Amon, "that I think would prepare us for this future."
