Qualcomm Revs Up AI Capabilities for Snapdragon Laptops and Tablets

pcmag.com:

Snapdragon 8cx Gen 3, Qualcomm's compute platform, isn't new—we first saw it in a laptop in early 2022—but the chip is back in the spotlight as Qualcomm rolls out new AI features and capabilities at Microsoft Build 2023. Partnering with Microsoft, the chip maker introduced its latest developments in on-device AI, including new ways for developers to leverage those capabilities on Snapdragon-powered devices.

Qualcomm's big bet: Edge devices—meaning the laptops, tablets, and phones in our hands—can work in tandem with server-based AI to dramatically improve the performance of different AI tools and services.

This is especially helpful in thin-and-light systems. Where dedicated GPUs are scarce, this added horsepower allows for on-device AI in new scenarios. On-device AI use promises to make generative AI tools, like image generation in Stable Diffusion, more accessible and affordable by shifting some of the workloads to your device. Local processing also has the potential to make these tools more reliable as well as improve personalization and privacy by allowing more data to be kept and used locally.

Now, Qualcomm just showed off its own AI stack, demonstrating how much better AI tools can be when run locally with Qualcomm's Hexagon processor on the Snapdragon 8cx Gen 3. And, to spur further development of these capabilities, the company will open up the Qualcomm AI Engine Direct SDK to the public, so that any developer can make the most of this hardware acceleration.

We recently sat down with Qualcomm to see these AI capabilities in action. The demo emphasized the capabilities of Qualcomm's Hexagon processor and the Qualcomm AI Engine, which leverages a separate bit of hardware for faster AI tasks, but without the