Microsoft Corp. today continued to advance its commitment to artificial intelligence in its Windows 11 operating system with an announcement during its Ignite 2023 conference unveiling updates to the developer experience and new features for its cloud-based Windows 365 service.

With the latest Windows 11 update on Sept. 26, Microsoft released a raft of developer features aimed at bringing AI as a core component of the OS intended to make the developer experience more productive. Now the company said is opening up new tools that will help developers jumpstart their AI ambitions on Windows to tackle and deploy conversational experiences.

Check out our full coverage of Microsoft Ignite in these stories:

• First dedicated AI accelerators and in-house Arm processors unveiled for Microsoft Azure
• Microsoft turbocharges AI development with Fabric, Azure AI Studio and database updates
• AI-powered Microsoft Copilot expands with new features and customization
• Broad collection of security products announced at Microsoft Ignite
• Microsoft debuts new AI tools for Windows developers and IT professionals
• Nvidia announces generative AI foundry on Azure and new foundational models
• Microsoft brings generative AI copilots to sales and service teams
• Exclusive: Informatica and Microsoft team on native data integration for Fabric analytics

The new Windows AI Studio will simplify AI app development by providing access to AI development tools and models from Azure AI Studio, also announced at Ignite, and also models from other distributors such as Hugging Face Inc., which will allow developers to build, fine-tune and deploy small language models for local use.

Small language models, or SLMs, as opposed to large language models such as OpenAI LP’s GPT-4, are lightweight and tailored to specific industry-focused or business needs. By contrast, LLMs are designed to be more generalized, take up vast amounts of processor cycles and are very costly to use, so they reside in the cloud and run on massive arrays of graphical processing units. Using an SLM also means all data processing happens locally, which means no data leaves, making them more secure.

The AI Studio provides developers with a guided workspace setup for model configurations and a user interface to help them fine-tune popular SLMs, such as Phi, Llama 2 and Mistral. From there, developers can rapidly iterate their model and fine-tune their model using Prompt Flow and Gradio templates from within the workspace.

A Visual Studio Code extension will also be rolling out in the next few weeks to make integrating models into Windows apps easier for developers looking to get started on AI app development. The guided interface that comes with the workspace will help developers focus on the development process, while the Studio itself does all the behind-the-scenes work of setting up the environment and provides the AI tools.

In addition to fine-tuning capabilities, AI Studio will also deliver a catalog of state-of-the-art models optimized specifically for Windows GPUs in the future, starting with Llama2-7B, Mistral-7B and Stable Diffusion XL.

This news comes on the cusp of Microsoft rolling out further updates to its Copilot AI-assistant, a powerful generative AI broadly integrated into Windows 11, the company’s Edge web browser and Bing search engine. The assistant is capable of providing conversational question-and-answer capabilities for users, helping automate tasks in Windows, writing documents in Word, generating complex calculations in Excel and more.

With access to the ability to create and deploy small language models locally, developers will be able to build custom AI-powered apps that can provide users conversational capabilities to also answer specific questions, automate workflows and solve problems within their specific contexts.

New features for Windows 365 and Azure Virtual Desktop

Today Microsoft also announced a number of new features and enhancements for Windows 365, its cloud-based OS service for Windows 11 or Windows 10, and Azure Virtual Desktop, which will make it easier for administrators to deploy desktops and systems to workers both in offices and remote.

Windows 365 provides what Microsoft calls “Cloud PCs,” or the ability to launch a fully complete and customized operating system from anywhere onto any machine, including MacBooks, iPads, Linux devices or Android phones. This allows users to pick up from where they left off on another Windows 365 instance and immediately access their data and desktop. Similarly, a Dev Box provides developers with self-service project-specific developer environments from the cloud with all the tools they need that they can spin up on any device.

Azure-based Windows Virtual Desktops are a slightly different approach to the same idea. They provide virtualized applications and remote desktops with support for Windows 10 and Windows 11 with personalized and multisession desktops.

Using Windows App, admins will be able to connect any device or app that can deploy Windows 365, Azure Virtual Desktop, Remote Desktop Services, Dev Box or other services. Using this system, a user can quickly pick their associated device such as a Cloud PC, Dev Box, or virtual desktop and then launch their favorite apps from a centralized application. Windows App is in public preview.

Also in preview is graphical processor unit support for Windows 365 for creative workloads such as graphic design, image and video rendering, 3D modeling, data processing and visualization. In preview as well are AI capabilities for Cloud PC resources, which can offer recommendations to help reduce costs, increase efficiency and aid with security and management.

On the security side, single-sign-on and passwordless authentication support for Windows 365 and Azure Virtual Desktop is generally available. Administrators can also block screen captures and protect against tampering and add watermarks to screens to assist with protecting proprietary information. A lockbox capability is in public preview that allows customers to lock up data and content so that Microsoft Support engineers cannot access it without explicit approval.

Coming into public preview soon, organizations will be able to encrypt their Windows 365 Cloud PC instance disks with their own encryption keys with managed keys. This will allow security professionals to maintain secrecy and security by controlling their data while their Cloud PC disks are at rest in the cloud, making it even harder for potential snoops to steal the information.

Image: Microsoft