Qualcomm sees AI as instrumental to its expansion beyond smartphones
Editor’s note: Qualcomm provided travel, lodging and other accommodations associated with the Snapdragon Tech Summit.
KONA, HAWAII–“The concept is this,” Qualcomm VP of Technology Jeff Gehlhaar told me in the press lounge at the Fairmont Orchid. “We see AI as a tool that is being used across industries to improve users’ lives. I’ll take the broadest brush. We’re involved in, say, automated driving. The purpose of automated driving…is to save the lives of pedestrians and passengers by reducing the error rate or, ideally, moving the error rate to zero for incidents. These are the kind of things that AI enables.”
To the notion of AI as becoming ubiquitous and democratized, words that are simultaneously easy to understand and incredibly difficult to contextualize, “What we really mean is we’re providing a horizontal portfolio of products involving hardware and software. We feel that Qualcomm is uniquely positioned to provide a ubiquitous position in hardware and software.”
This week at the Snapdragon Tech Summit, Qualcomm announced the Snapdragon 8 Gen 1 mobile platform, including its seventh generation AI Engine which boasts a 4x performance increase compared to its predecessor, and up to a 1.7x increase in performance per Watt. In terms of mobile user experiences, AI brings across the board benefits–better pictures, improved audio, increasingly robust security, more immersive gaming experiences, and on and on.
Beyond mobile, Qualcomm recently laid out an ambitious growth strategy—from $100 billion to $700 billion in the next decade. The idea is to leverage domain expertise developed for mobile devices and drive it into new areas, including automotive, the internet of things, gaming, PCs, extended reality, the metaverse, and more. All categories with clear applications for AI.
Gehlhaar has been with Qualcomm for around three decades. Reflecting on the pace of AI development over the most recent of those decades, he said, “I don’t know that we understood where it was going to go, but right away we could see applications in face detection, in object detection. We thought AI was going to go in the direction of neuromorphic computing…and then this what I’ll call the modern wave of what was at the time artificial neural networks emerged.”
For consumers, whether they’re aware of it or not, AI is instrumental in mobile photography. You don’t buy a new device and immediately upskill your photographic prowess; rather, the on-device AI is applying a range of tools to the image to make it look better. (That’s a highly reductive description of what is an unbelievably complex and nuanced process.) This raises an interesting question that generated a moderate amount of debate in an incredibly niche part of tech Twitter: are we really even taking pictures or has the combination of AI and ISPs using the picture you took and turning it into an AI-generated image based on the initial input?
“Taking a picture is an artistic expression,” Gehlhaar said. “AI is going to make that experience better but it’s also going to give you the power to do things like replicate experiences. I think that’s part of the artistic capture. It doesn’t change the nature of the image but, in some sense, it’s not the image you saw with your naked eye.”
To wildly zoom out, some in the tech industry see AI as inevitable. From this fatalistic point of view, the best bet is to just sit back and let the world of ubiquitous AI develop unimpeded. Others see us as at a point in time where AI-facing rules, frameworks and governance structures can still be developed and deployed. In a topical example, notable tech fiend Henry Kissinger, former Google CEO Eric Schmidt and MIT computer scientist Daniel Huttenlocher wrote in the book The Age of AI: And Our Human Future, “Without significant fanfare—or even visibility—we are integrating nonhuman intelligence into the basic fabric of human activity…The nations of the world must make urgent decisions regarding what is compatible with concepts of inherent human dignity and moral agency.”
I bounced that quote off of Gehlhaar. “I don’t think that AI is inherently sort of good or bad,” he said. For Qualcomm, “The kinds of use cases that we are enabling with Snapdragon 8 Gen 1 are use cases around image quality enhancement, noise suppression, productivity enhancements. These are not use cases of agency or control. We’re using AI and our partners…are using AI to enhance experiences that have already existed in the device and we can make them better with AI.”
