This video features a discussion between Nick (Qualcomm AI Developer Advocate) and Jeremy Sinclair (Snapdragon insider) from Microsoft Build, exploring key announcements and insights from the conference. They delve into surprising developments like an AI solution assisting the deaf community, the open-sourcing of Windows Subsystem for Linux (WSL), and advancements in on-device AI capabilities such as AI Foundry and native audio production tools running on NPUs. Other topics include the significance of Docker support on ARM, the strategic Qualcomm-Nvidia data center partnership, emerging enterprise AI applications, and the open-sourcing of GitHub Copilot. Jeremy emphasizes the growing interest and robust performance of ARM-based Windows devices, highlighting their efficiency and capabilities for developers.
Key Takeaways from Microsoft Build:
0:08 Microsoft Build Conference Overview: Nick (Qualcomm AI Developer Advocate) and Jeremy Sinclair (Snapdragon insider) discuss their experiences and key takeaways from the Microsoft Build conference.
1:14 AI for Good - Assisting the Deaf Community: Jeremy's "coolest thing" seen was a Japanese company using Azure OpenAI services to translate speech to text for deaf individuals, showcasing AI's potential for assistive technology and social good.
2:53 Windows Subsystem for Linux (WSL) Open-Sourced: The most surprising announcement for Jeremy was the open-sourcing of WSL, a significant move by Microsoft that will allow greater community involvement and foster seamless Linux application integration on Windows.
4:55 On-Device Audio Production with One Take Audio: A VST plugin from One Take Audio impressed Jeremy by using neural networks and voice models to change vocal pitch and voice (male/female) in real-time, running efficiently on the NPU (Neural Processing Unit). This signals a major step towards native Digital Audio Workstation (DAW) support on Snapdragon devices.
7:20 Docker Support on ARM: Official Docker support for ARM devices is highlighted as a crucial development for developers, enabling smoother workflows, containerization, and local development on edge devices.
10:18 AI Foundry for Local AI: The video discusses AI Foundry (formerly Windows Copilot Runtime), which allows AI models to run directly on devices, enhancing privacy and security by keeping data local. Jeremy plans to use it for automated alt-text generation for social media images.
13:27 Qualcomm-Nvidia Data Center Partnership: A surprising announcement revealed a partnership between Qualcomm and Nvidia for data centers, with Qualcomm providing CPUs and Nvidia supplying GPUs, marking a strategic collaboration between traditional competitors in the AI space.
15:14 Enterprise AI Adoption: AI is becoming increasingly useful in enterprise for document processing, data aggregation, summarization, and voice transcription, enhancing efficiency. Emerging multi-layered agentic AI systems are also discussed as a future trend.
18:50 GitHub Copilot VS Code Extension Open-Sourced: The GitHub Copilot VS Code extension is now open source, reflecting Microsoft's commitment to open source and enabling developers to switch between various underlying AI models (e.g., GPT-4o, Claude Sonnet, Gemini) for varied code generation.
20:56 Prism and ARM Native Application Performance: The discussion touches on Prism, the x86 emulation layer for ARM, noting that its performance is already impressive, with x86 apps running surprisingly well on ARM devices, even surpassing x86 machines in some cases, especially on the Windows Canary branch.
22:45 Jeremy's Role as an ARM Development Expert: Jeremy, as a Microsoft expert, is assisting developers at Build with porting applications to run natively on ARM. He notes a significant shift in developer interest towards ARM-based Windows devices ("Copilot+ PCs") due to their performance, efficiency, and exceptional battery life, even under heavy development workloads.
26:00 Call to Action for Developers: The video concludes with an invitation to a developer networking event at Build and a strong encouragement for developers to continue building native ARM applications to achieve a seamless "just works" Windows experience, regardless of underlying architecture.