Windows boss sees AI bridging cloud and PCs
Artificial intelligence (Windows AI) is transforming the way we interact with our devices, from smartphones to laptops. AI allows us to use natural language, voice, gestures and even eye movements to control our digital tools. But AI is not only changing the user interface, it is also improving the performance, security and productivity of our devices.
One of the key challenges of AI is making it work seamlessly across different platforms and environments. How can we ensure that our AI applications and services are consistent and compatible whether we use them in the cloud, on our PCs, or on our mobile devices? How can we harness the power of the cloud to increase the capabilities of our local devices and vice versa?
This is the vision of Panos Panay, product director of Windows and Microsoft Devices. In a recent interview with The Verge, Panay shared his thoughts on how AI can bridge the gap between the cloud and the PC, and how Windows 11 is designed to enable this integration.
Panay explained that Windows 11 It is based on the idea of “hybrid computing,” meaning it can run on-premises and cloud-based applications with equal efficiency and flexibility. For example, Windows 11 supports Android apps through the Amazon Appstore, which are streamed from the cloud using Intel Bridge technology. This allows users to access a wide range of applications not available natively on Windows, without compromising the performance or security of their PCs.
On the other hand, Windows 11 also harnesses the power of local hardware to improve the cloud experience. For example, Windows 11 uses AI to optimize the battery life and network bandwidth of PCs when connected to the cloud. It also uses AI to improve the quality and reliability of video calls and meetings, reducing background noise, adjusting lighting, and stabilizing camera movements.
Panay said that AI is not just a feature of Windows 11, but a core part of its architecture. He said Windows 11 is designed to be “AI ready,” meaning it can adapt to different scenarios and contexts, and learn from user feedback and behavior. He also said that Windows 11 is “AI responsible,” meaning it respects user privacy and choice and follows ethical principles and guidelines.
Panay believes that AI is the future of computing and that Windows 11 is the first step in making this vision a reality. He said that he wants Windows 11 to be “the most open platform for developers and creators to create incredible experiences with AI.” He also said that he wants Windows 11 to be "the most personal and productive platform for users to enjoy the benefits of AI."
Microsoft has focused on cloud-based AI solutions, such as Microsoft 365 Copilot, which aim to increase the productivity of its users. However, some of its hardware partners, such as AMD, Intel, and Qualcomm, have been developing their own AI processors that can execute AI tasks locally on the PC. Is this a sign of a showdown between Microsoft and its allies?
Not necessarily, according to Microsoft's top Windows executive, who spoke at AMD's "Advancing AI" event, where the company unveiled the Ryzen 8040 series of mobile processors with enhanced AI capabilities. He said cloud AI and on-premise AI can complement each other, rather than compete.
This is an important question, because Microsoft has a large share of the Windows and Microsoft 365 markets, which together reach hundreds of millions of consumers and businesses. Microsoft also generates most of its revenue from its Azure cloud platform, which powers its artificial intelligence services. Microsoft charges $30 a month for Microsoft 365 Copilot, which uses AI in the cloud to help users with tasks like writing, editing, and research.
But AMD and its rivals also have a strong interest in promoting on-premises AI, which can offer faster performance, lower latency, and better privacy than AI in the cloud. AMD showed off several applications that take advantage of its AI processors, including Adobe Photoshop and Lightroom, and BlackMagic's DaVinci Resolve, which use AI for image and video editing. Microsoft itself also uses local AI for some of its Windows features, such as Windows Studio Effects, which uses AI to enhance video calls.
Therefore, it appears that cloud AI and on-premises AI are not mutually exclusive, but rather different options that can suit different needs and preferences. Microsoft and its hardware partners may not agree on everything, but they can still work together to deliver the best AI experience to their customers.
One of the challenges of AI is balancing the trade-offs between cloud computing and on-premises computing. Cloud computing offers more power and scalability, but also requires more bandwidth and latency. On-premises computing offers more privacy and responsiveness, but also has more limitations and costs. How can we achieve the best of both worlds?
That's the question Pavan Davuluri, the new corporate vice president of Microsoft's Windows and Devices division, addressed in a recent interview. He revealed that Microsoft is working on a “hybrid engine” that leverages on-premises AI and the cloud to deliver optimal performance and user experience.
According to Davuluri, the hybrid engine is an intelligent system that can dynamically allocate AI tasks between the cloud and the local device, based on the user's context and preferences. For example, if the user is offline or has a low-bandwidth connection, the hybrid engine can run AI functions locally using the hardware and software capabilities of the device. If the user is online and has a high-bandwidth connection, the hybrid engine can offload AI functions to the cloud, where it can benefit from more resources and data.
The hybrid engine is not only a technical solution, but also a strategic one. It allows Microsoft to differentiate itself from its competitors, which may rely more on the cloud or on-premises AI. It also allows Microsoft to offer more value and choice to its customers, who can enjoy faster, smarter and more personalized AI services across different devices and scenarios.
