Microsoft Build 2024 | Top Stories

Microsoft is once again bringing together the community of developers, software engineers, IT professionals, and businesses to launch the latest additions to its services and platforms. Microsoft Build is back and we have all the details.

*We will update this post with daily releases and news.

When and where Microsoft Build 2024 is being held

Last year, it has already regained its face-to-face essence and offers the opportunity to attend in person at the Seattle Convention Center.

However, given the great reception in recent years thanks to greater accessibility, the event will also be streamed online and free of charge to anyone who registers. The dates chosen are 21, 22 and 23 May.

Microsoft Build 2024 Sessions

This edition will be full of new innovative proposals and sessions ranging from the latest innovations in code development or challenges, to case studies from world experts.

The sessions will be led by Satya Nadella (CEO of Microsoft) and major industry figures and partners such as Kevin Scott, Scott Guthrie, and Rajesh Jha. In addition, they will be organized in several virtual modes, to connect attendees with Microsoft engineers, experts, and other presenters during the meeting.

• One-on-one consultations: you will be able to meet with a Microsoft engineer to discuss specific questions and solutions.
• Microsoft Learn Cloud Skills Challenge: hone, expand, and discover new skills and get a free certification exam by completing any of the 8 unique Microsoft Learn collections (.NET Challenge, Azure AI Challenge, Cloud Development, Microsoft 365, Azure Cosmos DB, DevOps…).
• Imagine Cup: Millions of student developers participate in this global competition where they learn and make a difference in their communities.
• On-demand sessions: if you missed a session, don’t worry because you can watch it again thanks to the complete catalog of on-demand sessions.

Microsoft Build 2024 Recap

Summary Day 1

The kick-off of the first day of Microsoft Build 2024 has been full of new announcements, especially related to generative artificial intelligence, data, and security. Here is the Keynote given by Satya Nadella.

Copilot Studio

A new set of features was introduced in Copilot Studio that will enable the creation of copilots that can act as agents, execute and orchestrate many common business processes, integrate line-of-business data directly into copilot experiences, and publish new extensions.

They will bring together connectors from Microsoft Graph and Power Platform, AI skills in Microsoft Fabric, and Dataverse to make the process of seating co-pilots on proprietary and third-party line-of-business data a wizard-based experience configured through Copilot Studio.

Azure AI Search

This service is a full-featured information retrieval system built to run enterprise searches and enhanced retrieval generation.

New enhancements have now been introduced, such as:

• Performing RAG at scale with more capacity.
• More relevant results through vector enhancements and hybrid search.
• New data integrations and seamless processing with integrated vectorization upgrades.
• New platform integration with the OneLake connector for archives.

Foundation Models and GPT-4o

Key AI updates were introduced to deliver more customizable, efficient, and flexible applications.

• GPT-4o was released a few days ago and is now generally available in Azure AI Studio and as an API. It’s a groundbreaking multimodal model integrates text, vision, and audio capabilities, setting a new standard for generative and conversational AI experiences.
• GPT-4 has been fine-tuned and enhanced with unprecedented customization of AI models, ensuring results are closely aligned with the specific needs of organizations.
• Assistants API paves the way for the creation of advanced virtual assistants and chatbots that enhance user interactions with their nuanced understanding and responsiveness.

Real-Time Intelligence in Microsoft Fabric

Microsoft Fabric’s new Real-Time Intelligence will provide a comprehensive software-as-a-service (SaaS) solution that enables customers to act on large volumes of time-sensitive, highly granular data in a proactive and timely manner to make faster, more informed business decisions.

Already in preview, it will allow users to access simple, codeless experiences for everyday analysts, or code-rich interfaces for professionals:

• Real-Time Hub will be a single place to ingest, process, and route events into Fabric as a central point for managing events from a variety of sources.
• Event streams that will provide out-of-the-box stream connectors to cross-cloud sources and content-based routing that helps eliminate the complexity of ingesting stream data from external sources.
• Real-time event house and dashboards with enhanced data exploration to help business users looking to gain insights from terabytes of streaming data without writing code.
• Data Trigger will integrate with the real-time hub to make it easy to trigger any pattern or change in the data in real-time.
• AI-powered insights, now with an integrated Microsoft Copilot experience in Fabric for query generation, preview, and one-click anomaly detection experience.
• Event-driven Fabric will allow users to respond to system events that occur within Fabric and trigger actions, such as executing data pipelines.

GitHub Copilot

New GitHub Copilot extensions were introduced that allow you to customize your experience with preferred services such as Azure, Docker, Sentry, and more.

Thanks to GitHub Copilot Extensions, you’ll be able to build and deploy in the cloud in natural language with your chosen tools and services, allowing you to stay in the flow longer, improve your skills, and innovate faster.

From GitHub Copilot Chat, you can now easily invoke all these tools to get context, perform actions, and generate files and pull requests, accelerating workflows on more tools.

Copilot + PCs

These are the most intelligent PCs ever built, with AI-infused at every layer, powered by Neural Processing Units (NPUs) capable of delivering 45 TOPS of compute.

This new class of devices is an invitation to app developers to deliver differentiated AI experiences that run on the edge taking advantage of the NPUs to bring great benefits, minimal latency, cost efficiency, data privacy, and more.

Phi-3 Vision

Still, in the preliminary version, it is the new multimodal model of the Phi-3 family of small AI linguistic models.

Phi-3-vision supports general visual reasoning tasks as well as reasoning with charts, graphs, and tables. The model allows for inputting images and text and obtaining textual responses.

Power Automate Updates

Innovations have been launched in Power Automate that will capitalize on process automation and generative AI to help revolutionize the way people work:

• AI flows will offer a new type of outcome-based automation that will allow generative AI to achieve an outcome by creating a flow autonomously, without the need for users to plan each step in the process.
• Two new features are added to desktop flows, including a new recorder that will create a more natural multi-modal experience and natural language-to-flow conversion features for the desktop.

Windows Copilot Runtime

To help developers with new AI capabilities, Copilot Runtime extends Microsoft Copilot Stack to Windows and has AI infused into every layer of Windows, including a fundamental transformation of the operating system itself.

It will help developers build great AI experiences, regardless of where they are on their AI journey, including Copilot Library and AI frameworks and toolchains.

New Virtual Machines ND MI300X V5

Powering these technological advances requires more powerful computers with purpose-built AI accelerators that have resources such as high-bandwidth memory (HBM), specialized data formats, and exceptional computing performance.

To meet that need, Azure becomes the first cloud service to offer general availability of a new ND MI300X v5 virtual machine (VM) series based on AMD’s latest Instinct GPU, MI300X. This new VM series is the first cloud offering of its kind and is designed to deliver the highest bandwidth memory (HBM) capacity of any VM available with industry-leading speeds, enabling customers to serve larger models faster and with fewer GPUs.

Azure Cobalt

Also part of Azure’s new series of virtual machines optimized for AI and cloud-native workloads, Cobalt 100 Arm, the first generation of virtual machines featuring the new Cobalt 100 processor, custom-built on an Arm architecture optimized for efficiency and performance when running general-purpose and cloud-native workloads, has been introduced.

These virtual machines will offer users both performance consistency and linear performance scaling with workloads such as web applications, microservices, and open-source databases.

Microsoft Azure Infrastructure

They ensured they had the most scalable and comprehensive infrastructure in the market in the AI era, with more than 60 data centers worldwide, increasing capacity and investment across all continents.

They also introduced Azure Compute Fleet, a new service that will simplify the provisioning of Azure compute capacity across different virtual machines, availability zones, and pricing models to help customers achieve the scale, performance, and cost they need.

Better defense with Microsoft Edge

The service improves its defense against data leaks and vulnerabilities with two new features:

• Screenshot prevention protects against data leakage in the browser by blocking the ability to take screenshots on pages labeled as sensitive or protected.
• Easier updates: will allow IT administrators to see which devices have out-of-date and at-risk instances of Edge. It will also provide mitigation controls, such as restarting the browser, triggering updates automatically, or enabling enhanced security mode.

Responsible AI enhancements

To help ensure the safe and responsible use of AI, including Prompt Shields, and Groundedness Detection, are now available. Custom Categories are coming soon so developers can create custom filters for generative AI apps tailored to specific content filtering needs. Additionally, Custom Categories provides options for standard and rapid deployment, allowing users to quickly address incidents and emerging threats by deploying new filters in less than an hour.

Prompt Shields detect and block prompt injection attacks (users manipulating prompts to inject harmful inputs) and with Groundedness Detection, developers can detect and block “hallucinations” in model outputs. The new capabilities underscore our continued commitment to advancing responsible AI practices and equipping customers with advanced safeguards to protect their AI applications.

Microsoft Intelligent Data Platform

To help lay the foundation for innovating with AI, Microsoft is introducing a new way to manage data at any scale, anywhere. Where on-premises investments and experiences extend to the cloud, while managed services do more with modernized workloads for the next stage of growth.

By leveraging the fully managed capabilities of Azure databases and integration with Microsoft’s own and third-party solutions, developers can focus on what matters: innovation.

Summary Day 2

Copilot in SharePoint

Soon companies will be able to elevate employee self-service by allowing employees to create a co-pilot that provides a level of information, context, and discoverability from information extracted from documents and files hosted on SharePoint.

These will help the team find the files they need and answer questions based on site content. They are easy to create in a few clicks, with additional enhancements, allowing them to be shared within an organization.

AKS Automatic

Launches Automatic to facilitate Kubernetes adoption for developers, DevOps teams, and platform engineers. It will automate the configuration and management of AKS clusters, incorporating best-practice configurations, so users are assured of the security, performance, and reliability of their applications.

Automatic will provide access to Kubernetes APIs, preserving the flexibility of Kubernetes that is important to many customers:

• Option to apply deployment safeguards.
• Fleet Manager will enable intelligent placement of workloads by customizing and overriding specific cluster resources.
• Kubernetes Event Driven Autoscaler (KEDA) in the Azure portal will streamline the creation and management of KEDA resources through the portal interface.
• Automated deployments.
• Auto-provisioning of nodes.

Technology with Social Impact

We were able to discover some of the socially meaningful technology solutions created with generative AI, including a chatbot solution that helps those facing food shortages locate help more easily and AI voice technology that brings greater personalisation to text-to-speech solutions for ALS sufferers.

PyTorch in Windows

DirectML is now compatible with another popular framework, PyTorch. This will allow developers to bring Hugging Face models to Windows and, thanks to DirectML, developers can scale their AI innovation on a variety of Windows hardware.

New LLMs and SLMs

They will be expanding their selection of frontier and open LLMs and SLMs, providing flexibility to choose based on business needs.

Phi-3-mini is now available via Models-as-a-Service (MaaS), so you can get up and running quickly and easily. Core42 JAIS and Nixtla TimeGen-1 models are also available via MaaS in preview. AI21, Bria AI, Gretel, NTT DATA, Stability AI, and Cohere Rerank models will be available soon.

New capacities in Azure Cosmos DB

The database designed for AI, which allows creators to build responsive and intelligent apps with real-time data, ingested and processed at any scale, has updates and new features:

• Integrated vector database capabilities – will incorporate vector indexing, allowing data and vectors to be stored together and kept in sync.
• Migration from serverless to provisioned accounts: The transition can be made seamlessly through the Azure portal or the Azure command-line interface (CLI).
• Cross-region disaster recovery: A cluster replica can be created in another region. This cluster replica will be continuously updated with data written to the primary region.
• Azure Cosmos DB Vercel integration: Developers building apps with Vercel can now easily connect to an existing Azure Cosmos DB database or create new Azure Try Cosmos DB accounts on the fly and integrate them into their projects.
• Go SDK: Allows customers to connect to an Azure Cosmos DB account for NoSQL and perform operations on databases, containers, and elements.

Stay tuned as we continue to recap the highlights from day 3 of Microsoft Build 2024.