I constitute this blog each period to assistance navigate done the aggravated gait of quality and innovation we’re releasing for customers. And what a twelvemonth we’ve had! We witnessed unthinkable advances astatine breathtaking velocity arsenic AI reshaped what’s imaginable crossed industries, organizations, and our day-to-day lives.

Personally, I’m loving the features and capabilities of Copilot for Microsoft 365, particularly gathering summaries aft a Teams telephone truthful I tin rapidly entree the astir important points and immoderate actions. We’re inactive experimenting with AI to trade selling narratives crossed my squad and it’s breathtaking to spot however AI enhances our ain creativity.

AI exertion is not marque new. It’s been improving experiences crossed applications and concern processes for immoderate clip but, the wide availability of generative AI models and tools this twelvemonth was a watershed moment.  

Customers and partners made speedy pivots to bring AI into their translation roadmaps, and wrong months were deploying applications and services powered by AI. I person ne'er seen a caller exertion make specified profound alteration astatine this pace. It shows however acceptable galore organizations were for this moment; however investments made successful the cloud, data, DevOps, and creating translation cultures acceptable the signifier for AI adoption. This twelvemonth Microsoft introduced hundreds of resources, models, services, and tools to assistance customers and partners maximize AI.

This month, I’m pleased to grow this blog’s scope and bring successful what’s caller for integer applications for a holistic look astatine everything we’re delivering to assistance customers modernize their information estate, physique intelligent applications, and use AI technologies to assistance execute their concern goals. Let’s go.

New models and multimodal capabilities available successful Azure AI

Our absorption is to present the astir cutting-edge unfastened and frontier models disposable truthful developers tin physique with assurance and unlock contiguous worth crossed their organization.  

Last period we announced a important enlargement of Azure OpenAI Service and introduced Models arsenic a Service (MaaS). This is simply a mode for exemplary providers to connection their latest unfastened and frontier LLMs connected Azure for generative AI developers to integrate into their applications.

And conscionable past week, we announced the availability of MaaS for Llama 2. With MaaS for Llama 2, developers tin integrate with their favourite LLM tools similar Prompt Flow, Semantic Kernel, and LangChain with a ready-to-use API and pay-as-you spell billing based connected tokens for LLMs. This allows generative AI developers to entree Llama 2 via hosted fine-tuning without provisioning GPUs, greatly simplifying the exemplary acceptable up and deployment process. Then they tin connection their customized applications utilizing Llama 2, purchased done and hosted connected the Azure Marketplace. See John Montgomery’s blog post connected this announcement for each the details and a look astatine respective different models disposable successful the Azure AI Model Catalog.  

Model catalog successful Azure AI Studio Preview

We proceed to beforehand our Azure OpenAI Service and precocious launched respective multimodal AI capabilities that empower businesses to physique generative AI experiences with image, substance and video, including:

• DALL·E 3, successful preview: Generate images from substance descriptions. DALL·E 3 is simply a singular AI exemplary that does conscionable that. Users picture an image, and DALL·E 3 volition beryllium capable to make it.
• GPT-3.5 Turbo exemplary with a 16k token punctual length, mostly available, and GPT-4 Turbo, successful preview: The latest models successful Azure OpenAI Service alteration customers to widen punctual magnitude and bring much power and ratio to their generative AI applications.
• GPT-4 Turbo with Vision (GPT-4V), successful preview: When integrated with Azure AI Vision, GPT-4V enhances experiences by allowing the inclusion of images oregon videos on with substance for generating substance output, benefiting from Azure AI Vision enhancement similar video analysis.
• Fine-tuning of Azure OpenAI Service models: Fine-tuning is present mostly disposable for Azure OpenAI Service models including Babbage-002, Davinci-002, and GPT-35-Turbo. Developers and information scientists tin customize these Azure OpenAI Service models for circumstantial tasks. Learn much astir fine-tuning here.
• GPT-4 updates: Azure OpenAI Service has besides rolled retired updates to GPT-4, including the quality for fine-tuning. Fine-tuning volition let organizations to customize the AI exemplary to amended suit their circumstantial needs. It’s akin to tailoring a suit to acceptable perfectly, but successful the satellite of AI. Updates to GPT-4 are successful preview.

Steering astatine the frontier: Extending the powerfulness of prompting

The powerfulness of prompting successful GPT-4 continues to amaze! A caller Microsoft Research blog discusses promptbase, an attack to prompting GPT-4 that harnesses its almighty reasoning abilities. Across a wide assortment of trial sets (including the ones utilized to benchmark the precocious announced Gemini Ultra), GPT-4 gets amended results than different AI models (including Gemini Ultra). And it does this with zero-shot chain-of-thought prompting. Read much successful the blog post and these resources are disposable connected GitHub to effort close now.

Infuse Responsible AI (RAI) tools and practices successful your LLMOps

As AI adoption matures and companies enactment their AI apps into production, it’s important to see the information boundaries supporting the short- and semipermanent ROI of those applications. This month, we continued our bid connected LLMOps for concern leaders with an nonfiction focused connected how to infuse liable AI into your AI improvement lifecycle. We item champion practices and tools successful Azure AI Studio to assistance improvement teams enactment their principles into practice.

Azure AI Advantage offer

Azure Cosmos DB is the unreality database for the Era of AI. It supports built-in AI, including earthy connection queries, AI vector hunt capabilities, and elemental AI integration with Azure AI Search. We’re investing successful helping customers observe these benefits with our caller Azure AI Advantage offer, which helps caller and existing Azure AI and GitHub Copilot customers prevention erstwhile utilizing Azure Cosmos DB by providing 40,000 Request Units per 2nd (RU/s) of Azure Cosmos DB for 90 days.

What’s caller successful the Azure information platform—because AI is lone arsenic bully arsenic your data

Every intelligent app starts with information truthful a modern information and analytics level is indispensable for immoderate AI transformation.

One illustration of this however multinational instrumentality steadfast Clifford Chance leverages caller technologies to payment their clients. The steadfast built a coagulated information level connected Azure to innovate with caller technologies, including the caller procreation of ample connection models, Azure OpenAI and Microsoft 365 Copilot. Early innovations are already delivering worth with cognitive translation proving to beryllium 1 of the fastest increasing products their IT squad has ever released.

And Belfius, a Belgian security company, built connected the Microsoft intelligent information level utilizing services similar Azure Machine Learning and Azure Databricks to trim improvement time, summation efficiency, and summation reliability. As a result, their information scientists tin absorption connected creating and transforming features and the institution tin amended observe fraud and wealth laundering. 

Azure and Databricks: Co-innovation for almighty AI experiences

Microsoft’s Scott Guthrie (left) and Arun Ulagaratchagan (right) with Databrick’s Co-Founder and CEO Ali Ghodsi (center) astatine Ignite

At Microsoft Ignite 2023 successful November the benefits of maturing AI tools and services were successful afloat presumption arsenic customers and partners shared however Microsoft is empowering them to execute more. 

Databricks is of 1 of our astir strategical partners with immoderate of the fastest increasing information services connected Azure. We precocious showcased our co-innovation, including Azure Databricks interoperability with Microsoft Fabric, and however Azure Databricks is taking vantage of Azure OpenAI to present AI experiences for Azure Databricks’ customers. This means customers tin instrumentality vantage of LLMs successful Azure OpenAI arsenic they physique AI capabilities similar retrieval-augmented procreation (RAG) applications connected Azure Databricks, and past usage Power BI successful Fabric to analyse the output.

If you missed Databricks’ CEO Ali Ghodsi connected signifier with Scott Guthrie astatine Ignite, I promote you to cheque retired the replay of Scott’s afloat conception Microsoft Cloud successful the epoch of AI. If you lone person a fewer minutes, his wrapper up connected LinkedIn is simply a large option, too. It’s a adjuvant overview of however the Microsoft Cloud is uniquely positioned to empower customers to alteration by gathering AI solutions and unlocking information insights utilizing the aforesaid level and services that powerfulness each of Microsoft’s broad solutions.

And for a person look astatine our latest enactment with Databricks, cheque retired the click-thru mentation of the Modern Analytics with Microsoft Fabric and Azure Databricks DREAM Lab from Ignite.

Now successful preview: Azure AI hold for Azure Database for PostgreSQL 

The caller Azure AI hold allows developers to leverage ample connection models (LLMs) successful Azure OpenAI to make vector embeddings and physique rich, PostgreSQL generative AI applications. These almighty caller capabilities combined with existing enactment for the pgvector extension, marque Azure Database for PostgreSQL different large destination for gathering AI-powered apps. Learn more. 

Azure Arc brings unreality innovation to SQL Server anywhere

This month, we’re introducing a caller acceptable of enhanced manageability and information capabilities from SQL Server enabled by Azure Arc. With Monitoring for SQL Server, customers tin summation captious insights into their full SQL Server property and optimize for database performance. Customers tin besides presumption and negociate Always On availability groups, failover clump instances, and backups straight from the Azure portal, with amended visibility and simplicity. Lastly, with Extended Security Updates arsenic a work and automated patching, customers tin ever support their apps secure, compliant, and up to day astatine each times. Learn more.

Lower pricing for Azure SQL Database Hyperscale compute

New pricing connected Azure SQL Database Hyperscale offers cloud-native workloads the show and information of Azure SQL astatine the terms of commercialized open-source databases. Hyperscale customers tin prevention up to 35% connected the compute resources they request to physique scalable, AI-ready unreality applications of immoderate size and I/O requirement. The caller pricing is present available. Learn more.

Digital applications present transformational operations and experiences

This epoch of AI is brought to beingness done integer applications developed and deployed by companies putting AI to enactment to heighten their operations and experiences—like a personalized app acquisition for employees oregon a customized chatbot for extremity customers. Here are immoderate caller updates that assistance marque each this innovation possible.

The 7 pillars of modern AI development: Leaning into the epoch of customized copilots

Copilots are generating a batch of excitement and with Azure AI Studio successful nationalist preview, developers person a level purpose-built for generative AI exertion development. As we thin into this caller era, it’s important for businesses to cautiously see however to plan a durable, adaptable, and effectual approach. How tin AI developers guarantee their solutions heighten lawsuit engagement? Here are 7 pillars to deliberation done erstwhile gathering your customized copilot.

AKS is simply a starring level for modern unreality autochthonal and intelligent applications

The aboriginal of app improvement is astatine the intersection of AI and cloud-native technologies similar Kubernetes. Cloud-native and AI are profoundly rooted unneurotic successful fueling innovation astatine standard and Azure Kubernetes Service (AKS) provides the standard customers request to tally their compute intensive workloads similar AI and instrumentality learning. Check retired Brendan Burn’s recent blog from KubeCon to larn however Microsoft is gathering and servicing open-source communities that payment our customers.

There are nary limits to your innovation with Azure

It has been astonishing to spot the tech assemblage and lawsuit effect to our caller news, resources, and features crossed the portfolio, peculiarly integer applications. 

Resources similar the Platform Engineering Guide launched successful with Ignite person been incredibly popular, demonstrating the request and appetite for this benignant of grooming material.  

Seeing however organizations innovate with the exertion is what it’s each about.   

Here are a mates of lawsuit stories that caught my attraction recently.

Modernizing interactive experiences crossed LEGO House with Azure Kubernetes Service

We’re collaborating with The LEGO House successful Denmark—the eventual LEGO acquisition halfway for children and adults—to migrate custom-built interactive integer experiences from an aging on-prem information halfway to Microsoft Azure Kubernetes Service (AKS) to amended stability, security, and the quality to iterate and collaborate connected caller impermanent experiences. This displacement to the unreality enables LEGO House to much rapidly update these experiences arsenic they larn from guests. As the destination modernizes it hopes to stock learnings and technologies with the broader LEGO Group ecosystem, similar LEGOLAND and marque retail stores. 

Gluwa chose Azure for a reliable, scalable unreality solution to bring banking to emerging, underserved markets and adjacent the fiscal gap

An estimated 1.4 cardinal radical deficiency entree to basal fiscal services due to the fact that they unrecorded successful a state with constricted fiscal infrastructure, making it hard to get recognition oregon idiosyncratic and concern loans. Gluwa with Creditcoin uses blockchain exertion to differentiate their concern done borderless fiscal exertion and chose Azure arsenic the instauration to enactment it. Using a wide operation of our services and solutions—.NET framework, Azure Container Instances, AKS, Azure SQL, Azure Cosmos DB, and more—Gluwa has a beardown level to enactment their offerings. The concern has besides boosted operational ratio with reliable uptime, unchangeable services, and affluent merchandise offerings. 

CARIAD creates a work level for Volkswagen Group vehicles with Azure and AKS

With the automotive manufacture shifting to bundle defined vehicles, CARIAD, the Volkswagen Group bundle subsidiary, collaborated with Microsoft utilizing Azure and AKS to make the CARIAD Service Platform for providing automotive applications to brands similar Audi, Porsche, Volkswagen, Seat and Skoda. This level powers and accelerates the improvement and work of conveyance bundle by CARIAD’s developers, helping bundle go an vantage for the Volkswagen Group successful the adjacent procreation of automotive mobility.

DICK’S Sporting Goods creates an omnichannel lawsuit acquisition utilizing Azure Arc and AKS

To make a much consistent, personalized acquisition for customers crossed its 850 stores and its online retail experience, DICK’S Sporting Goods envisioned a “one store” exertion strategy with the quality to write, deploy, manage, and show its store bundle crossed each locations nationwide—and bespeak those aforesaid experiences done its eCommerce site. DICK’S needed a caller level of modularity, integration, and simplicity to seamlessly link its nationalist unreality situation with its computing systems astatine the edge. With the assistance of Microsoft, DICK’s Sporting Goods is migrating its on-premise infrastructure to Microsoft Azure and creating an adaptive unreality situation comprised of Azure Arc and Azure Kubernetes Service (AKS). Now the retailer tin easy deploy caller applications to each store to enactment a ubiquitous experience. 

Azure Cobalt delivers show and ratio for intelligent applications

Azure provides hundreds of services supporting the show demands of unreality autochthonal and intelligent applications. Our enactment to maximize show and ratio present extends to the silicon powering Azure. We precocious introduced Azure Maia, our archetypal customized AI accelerator bid to tally cloud-based grooming and inferencing for AI workloads, and our customized in-house CPU series, Azure Cobalt, the archetypal CPU designed by us, specifically for the Microsoft Cloud. 

Cobalt 100, the archetypal procreation successful the series, is simply a 64-bit 128-core spot that delivers up to 40% show betterment implicit existent generations of Azure Arm chips and tin powerfulness services specified arsenic Microsoft Teams and Azure SQL.

With Cobalt, we volition present show and economics to conscionable the demands of resource-intensive workloads similar intelligent applications utilizing generative AI for at-scale workloads. Our customized silicon and strategy features see dynamic powerfulness capabilities that tin beryllium tuned per-core based connected the workload, leveraging the co-optimization of hardware with bundle to present best-in-class show efficiency.  Check retired Omar Khan’s blog, Microsoft Azure delivers purpose-built unreality infrastructure successful the epoch of AI, for much connected however we’re redefining unreality infrastructure from systems to silicon to maximize the epoch of AI.  

Don’t conscionable instrumentality it from maine and Omar. Rani Borkar, CVP, Azure Hardware Systems shares however Cobalt delivers differentiated performance—and tin adjacent assistance thrust toward achieving sustainability goals:

“We’re gathering sustainability into each portion of our hardware for the cloud, from our silicon to the servers. This starts successful the plan phase, and connected Cobalt we made those intentional plan choices to beryllium capable to power show and powerfulness depletion per halfway and connected each azygous VM. We’re pleased with the show we’ve seen investigating Cobalt connected interior workloads similar Teams and Azure SQL, and looking guardant to rolling this retired much wide to customers arsenic a VM offering adjacent year.”

Grow your data, AI, and intelligent applications knowing and skillset

I often usage December downtime to bash a heavy dive into a caller exertion oregon larn a caller skill. I’m a information geek, and I person fond memories from a fewer years agone erstwhile I worked done much than a twelve Power BI trainings and loved each infinitesimal of it. Demystifying AI and helping customers physique the close skillsets is conscionable 1 mode Microsoft is empowering AI transformation. Here are immoderate caller learning resources if you are similar maine and program to extremity the twelvemonth learning thing new!

AI successful a Minute: Looking for assistance ramping your teams connected generative AI? Or possibly you privation to spell to extremity of twelvemonth gatherings prepared to sermon AI and its possibilities? Our caller “AI successful a Minute” video series explains generative AI basics successful short, snackable bites anyone tin digest, careless of occupation tile, level, oregon industry.

GenAI for beginners: This free 12-lesson course is designed to thatch beginners everything they request to cognize to commencement gathering with generative AI. 

Microsoft Learn Cloud Skills Challenge—Microsoft Ignite Edition: Skill up for in-demand Data & AI tech scenarios and participate to triumph a VIP walk to the adjacent Microsoft Ignite oregon Microsoft Build by completing a situation by January 15, 2024. 

Cloud Workshop for the SQL Professional—If you missed this store astatine the PASS Data Summit, the labs are disposable to spell done astatine your ain pace. Our adjacent store volition beryllium successful March astatine the Microsoft Fabric Community Conference, truthful beryllium definite to registry and travel spot america successful Las Vegas.

Accelerate your AI travel with cardinal solution accelerators—Get started gathering intelligent apps connected Azure with newly-published demos, GitHub repos, and Hackathon contented to make AI-powered, intelligent apps. Customers and partners tin “Build Your Own Copilot” utilizing a click-through demo, sellers and partners tin tally a Hackathon with customers, and developers tin leverage a codification repo to rapidly physique solutions successful their Azure account. Get started here: Build Modern AI Apps, Hackathon, Vector hunt AI Assistant.

What volition you physique successful 2024? Transform your concern with a trusted partner

The quality of AI to accelerate translation crossed industries, organizations, and regular beingness volition surely proceed astatine an aggravated gait successful 2024. Microsoft is arrogant to beryllium your trusted spouse successful this epoch of AI and we’re committed to helping you execute much for your business. I americium excited to spot however data, AI, and integer applications innovation unfolds for your concern successful the caller year!

The station What’s caller successful Azure Data, AI, & Digital Applications: Modernize your information estate, physique intelligent apps, and use AI solutions appeared archetypal connected Microsoft Azure Blog.