Justin Farris, a longtime product leader at Zillow and GitLab, joined Seattle startup Read AI as vice president of product.

Farris spent more than five years in a similar role at GitLab, the publicly traded DevOps company, and was director of product development at Zillow.

Founded in 2021, Read AI got its start with a meeting notetaker and has added various other enterprise productivity tools. It raised a $50 million Series B round last year.

In a press release, Farris called Read AI the “Cursor for productivity.”

“Just as GitLab is the go-to place for the entire software development process, Read AI is the go-to place where work comes together,” he said. “I saw first-hand what’s happening in the developer tools space, and see even larger opportunities in the productivity and AI assistant category.”

Read AI on Wednesday rolled out a new “Agentic Workflow Suite.” The company is led by David Shim, who won CEO of the Year honors at the GeekWire Awards in May.

— Lisa Gurry is stepping down from Seattle health data startup Truveta.

Gurry spent more than 23 years at Microsoft before helping launch Truveta as its chief marketing officer in 2020. She later served as chief operating officer and chief growth officer.

“I embraced my role as the only female co-founder and worked hard to create a successful company and a positive experience for all Truvetans,” Gurry wrote on LinkedIn.

Gurry said that she is “so excited for my next grand adventure” but didn’t reveal more details.

Truveta reached unicorn status earlier this year after raising $320 million. The company aims to aggregate medical records data from partner institutions to link treatments with outcomes and underlying health. It is led by CEO Terry Myerson, a former Microsoft executive vice president.

— Hossein Nowbar, chief legal officer at Microsoft, is stepping down after nearly three decades at the Redmond tech giant.

Nowbar graduated from the University of Washington in 1993, worked at Davis Wright Tremaine for four years, and then joined Microsoft in 1997.

He was named chief legal officer in 2023.

“I helped clear antitrust hurdles for transformative acquisitions like Nuance and Activision Blizzard King, supported product and service launches that reshaped the way we work, and introduced the Customer Copyright Commitment to help our customers realize the promise of AI,” Nowbar wrote on LinkedIn.

Nowbar didn’t share details of his next gig. Jonathan Palmer, a corporate vice president and 15-year veteran of Microsoft, will serve as the company’s new chief legal officer.

— Edwin Miller, CEO of Seattle-based Marchex, is stepping down after two years leading the publicly traded call and conversation analytics company. Marchex also announced a series of other exec changes:

• Troy Hartless, chief revenue officer, is also now president
• Francis Feeney, chief corporate and legal affairs officer, is also now chief operating officer
• Brian Nagle, senior vice president, controller, is now chief financial officer

Miller will serve as a senior advisor to Russell Horowitz, chairman at Marchex.

— Bridget Frey, a longtime leader at Seattle-based Redfin, is taking on the chief data officer role at Rocket after the mortgage giant completed its acquisition of Redfin in July.

Frey, who joined Redfin in 2011, will continue her role as chief technology officer at Redfin, where she helped launch the machine learning team.

“With Bridget guiding our combined data vision, we are set up to create a homeownership experience that feels genuinely connected end-to-end,” Rocket CTO Shawn Malhotra said on LinkedIn.

— Amazon Web Services has brought on two senior executives, according to a report from Business Insider.

• David Richardson, who was at Stripe and previously spent 16 years at AWS, is back at the company as vice president of AgentCore.
• Joe Hellerstein, a professor at UC Berkeley, is joining as vice president and distinguished scientist, according to Business Insider.

— Ty Wolfe-Jones joined Seattle-based EV charging startup Juicer Energy as head of operations. Wolfe-Jones was a director at Assurance, a COO at Wrench, and manager at both DoorDash and Uber.

Juicer, led by OfferUp co-founder Nick Huzar, recently started rolling out its EV charging infrastructure at apartment buildings and other properties.

“It is rare that you find big, real-world, ‘right now’ problems being dealt with in truly innovative ways, and with such a strong team,” Wolfe-Jones said in a statement. “So when you do, you have to jump at it.”

— Bryan Copley, a Seattle-based entrepreneur and founder of real estate data startup CityBldr, joined Blue Fern Development as chief product officer. Blue Fern, a Redmond, Wash.-based real estate development and construction firm, acquired CityBldr, according to Copley.

“I’m particularly excited about the opportunity to work with the team to pioneer combined learning in real estate development, fusing AI insights and human judgment to unlock greater productivity, which from this point forward will be measured in more homes,” he said on LinkedIn.

— Sarah Strobhar joined Seattle startup Upbound as chief revenue officer. Strobhar spent more than five years at AWS as a go-to-market leader and was also head of sales at Onera.

Founded in 2017, Upbound helps enterprise customers streamline their infrastructure workflow in the cloud. The company, founded and led by Bassam Tabbara, raised $60 million in 2021.

— Jennifer Adair, a longtime researcher at Seattle-based cancer center Fred Hutch, joined the UMass Chan Medical School as a professor and vice chair in the new Department of Genetic and Cellular Medicine. Adair spent 17 years at Fred Hutch, where she helped lead work on gene therapies at the Adair Lab. She was also an associate professor at Fred Hutch and a research associate professor at the University of Washington.

— Pablo Brown-Rodriguez, a former general manager at Amazon Health, launched a new Seattle-based firm called Operative Partners that acquires and leads mid-market healthcare businesses. “The goal is to find a great healthcare business and protect legacy, culture, and people with a long-term, operator-led approach,” he wrote on LinkedIn.

The home appliance industry would like you to believe that gas-burning stoves are not a risk to your health -- and several companies that make the devices are scrambling to erase their prior acknowledgements that they are. From a report: That claim is at the heart of a lawsuit the Association of Home Appliance Manufacturers has filed against the state of Colorado to stop it from requiring natural gas stoves, which burn methane, to carry health labels not unlike those on every pack of cigarettes. "Understand the air quality implications of having an indoor gas stove," the warning would read. The law was to take effect August 5 but is now on hold, and state officials did not respond to a request for comment. In its federal lawsuit, the Association -- whose board includes representatives of LG Electronics, BSH Home Appliance Corp. (which makes Bosch appliances), Whirlpool, and Samsung Electronics -- asserts that the labeling requirement is "unconstitutional compelled speech" and illegal under the First Amendment. It calls the legislation a climate law disguised as a health law and, most strikingly, it claims there is "no association between gas stoves and adverse health outcomes."

Read more of this story at Slashdot.

The next generation of Apple’s MacBook Pro laptops could be the first to feature a touchscreen display, according to supply chain analyst Ming-Chi Kuo. The change could apply to OLED MacBook Pro models that are expected to enter production by late 2026, Kuo reported on X, and will incorporate a “touch panel using on-cell touch technology.”

This builds on previous reports from Bloomberg’s Mark Gurman regarding Apple’s interest in developing touchscreen MacBooks, having said two years ago that the company may launch its first touch-panel Mac in 2025 “as part of a larger update to the MacBook Pro.” Gurman revisited these rumors in June, saying that he expects “future iPads and Macs to both have touch screens (with vibrant OLED displays), run the same apps, and have similar user interfaces.”

With the release of iPadOS 26, the iPad took a big step towards being a computer — maybe now it’s the MacBook’s turn to embrace the touchscreen interface an entire generation has been weaned upon. “This shift appears to reflect Apple’s long-term observation of iPad user behavior,” Kuo said, suggesting that adding touch controls could boost productivity for MacBook users in “certain scenarios.” I’m curious to see if future touchscreen MacBook Pro models will adopt Apple Pencil support or if that’s a step too far.

We’re happy to announce that SQL Server 2025 Release Candidate 1 (RC1) now includes preview support for Red Hat Enterprise Linux (RHEL) 10, expanding our commitment to modern, secure, and flexible Linux-based deployments.

RHEL 10 Support in SQL Server 2025 RC1

You can now deploy SQL Server 2025 Preview on RHEL10 for your Dev/Test environments using the Enterprise Evaluation Edition, which is valid for 180 days. For your production workloads you could use SQL Server 2022 on RHEL 9 or Ubuntu 22.04.

Deploying SQL Server 2025 RC1 on RHEL10

You can follow the Quickstart: Install SQL Server and create a database on RHEL10 to install SQL Server and create a database on RHEL10. It walks you through everything—from preparing your system to installing and configuring SQL Server.

To explore the latest improvements in SQL Server 2025 RC1, check out What's New in SQL Server 2025 - SQL Server | Microsoft Learn. I was particularly interested in testing the new Half-precision float support in vector data type.

To do this, I deployed SQL Server RHEL10 (the tag is 2025-RC1-rhel-10) container on WSL2 and I already have Docker Desktop installed on my local machine to manage containers. I launched the SQL Server 2025 RC1 container, connected to it using SQL Server Management Studio (SSMS), and successfully tested the vector data type enhancement. 

docker pull mcr.microsoft.com/mssql/rhel/server:2025-RC1-rhel-10 docker run -e "ACCEPT_EULA=Y" -e "MSSQL_SA_PASSWORD=passwordshouldbestrong" \ -e "MSSQL_AGENT_ENABLED=true" \ -p 14337:1433 --name sql2025RC1RHEL10 --hostname sql2025RC1RHEL10 \ -d mcr.microsoft.com/mssql/rhel/server:2025-RC1-rhel-10SELECT @@VERSION GO CREATE DATABASE SQL2025onRHEL10 GO USE SQL2025onRHEL10 GO -- Step 0: Enable Preview Features ALTER DATABASE SCOPED CONFIGURATION SET PREVIEW_FEATURES = ON; GO -- Step 1: Create a Table with a VECTOR(5, float16) Column CREATE TABLE dbo.Articles ( id INT PRIMARY KEY, title NVARCHAR(100), content NVARCHAR(MAX), embedding VECTOR(5, float16) ); -- Step 2: Insert Sample Data INSERT INTO Articles (id, title, content, embedding) VALUES (1, 'Intro to AI', 'This article introduces AI concepts.', '[0.1, 0.2, 0.3, 0.4, 0.5]'), (2, 'Deep Learning', 'Deep learning is a subset of ML.', '[0.2, 0.1, 0.4, 0.3, 0.6]'), (3, 'Neural Networks', 'Neural networks are powerful models.', '[0.3, 0.3, 0.2, 0.5, 0.1]'), (4, 'Machine Learning Basics', 'ML basics for beginners.', '[0.4, 0.5, 0.1, 0.2, 0.3]'), (5, 'Advanced AI', 'Exploring advanced AI techniques.', '[0.5, 0.4, 0.6, 0.1, 0.2]'); -- Step 3: Perform a Vector Similarity Search Using VECTOR_DISTANCE function DECLARE @v VECTOR(5, float16) = '[0.3, 0.3, 0.3, 0.3, 0.3]'; SELECT TOP (3) id, title, VECTOR_DISTANCE('cosine', @v, embedding) AS distance FROM dbo.Articles ORDER BY distance; -- Step 4: Optionally Create a Vector Index CREATE VECTOR INDEX vec_idx ON Articles(embedding) WITH ( metric = 'cosine', type = 'diskANN' ); -- Step 5: Perform a Vector Similarity Search DECLARE @qv VECTOR(5, float16) = '[0.3, 0.3, 0.3, 0.3, 0.3]'; SELECT t.id, t.title, t.content, s.distance FROM VECTOR_SEARCH( table = Articles AS t, column = embedding, similar_to = @qv, metric = 'cosine', top_n = 3 ) AS s ORDER BY s.distance, t.title;

Conclusion

The addition of RHEL10 support in SQL Server 2025 Preview is a major milestone in delivering a modern, secure, and flexible data platform for Linux users. We encourage you explore these new capabilities and share your feedback to help us continue enhancing SQL Server for the Linux ecosystem. 

You can share your feedback using any of the following methods: 

1. Email us at sqlpreviewpackage@microsoft.com with your thoughts and suggestions.
2. Submit your ideas on Azure Ideas (Use the SQL Server on Linux Group on the left side of the page)
3. Alternatively, you can open issues related to the preview packages Issues · microsoft/mssql-docker (github.com) on GitHub. 

We hope you give SQL Server 2025 preview on RHEL10 a try - and we look forward to hearing what you think!

What Is Foundry Local?

Foundry Local is a lightweight runtime designed to run AI models directly on user devices. It supports a wide range of hardware (CPU, GPU, NPU) and provides a consistent developer experience across platforms. The SDKs are available in multiple languages, including Python, JavaScript, Rust, and now C#.

Why a C# SDK?

The C# SDK brings Foundry Local into the heart of the .NET ecosystem. It allows developers to:

• Download and manage models locally.
• Run inference using OpenAI-compatible APIs.
• Integrate seamlessly with existing .NET applications.

This means you can build intelligent apps that run offline, reduce latency, and maintain data privacy—all without sacrificing developer productivity.

Bootstrap Process: How the SDK Gets You Started

One of the most developer-friendly aspects of the C# SDK is its automatic bootstrap process. Here's what happens under the hood when you initialise the SDK:

1. Service Discovery and Startup The SDK automatically locates the Foundry Local installation on the device and starts the inference service if it's not already running.
2. Model Download and Caching If the specified model isn't already cached locally, the SDK will download the most performant model variant (e.g. GPU, CPU, NPU) for the end user's hardware from the Foundry model catalog. This ensures you're always working with the latest optimised version.
3. Model Loading into Inference Service Once downloaded (or retrieved from cache), the model is loaded into the Foundry Local inference engine, ready to serve requests.

This streamlined process means developers can go from zero to inference with just a few lines of code—no manual setup or configuration required.

Leverage Your Existing AI Stack

One of the most exciting aspects of the Foundry Local C# SDK is its compatibility with popular AI tools such as:

• OpenAI SDK - Foundry local provides an OpenAI compliant chat completions (and embedding) API meaning. If you’re already using `OpenAI` chat completions API, you can reuse your existing code with minimal changes.
• Semantic Kernel - Foundry Local also integrates well with Semantic Kernel, Microsoft’s open-source framework for building AI agents. You can use Foundry Local models as plugins or endpoints within Semantic Kernel workflows—enabling advanced capabilities like memory, planning, and tool calling.

Quick Start Example

Follow these three steps:

1. Create a new project

Create a new C# project and navigate to it:

dotnet new console -n hello-foundry-local cd hello-foundry-local

2. Install NuGet packages

Install the following NuGet packages into your project:

dotnet add package Microsoft.AI.Foundry.Local --version 0.1.0 dotnet add package OpenAI --version 2.2.0-beta.4

3. Use the OpenAI SDK with Foundry Local

The following example demonstrates how to use the OpenAI SDK with Foundry Local. The code initializes the Foundry Local service, loads a model, and generates a response using the OpenAI SDK.

Copy-and-paste the following code into a C# file named Program.cs:

using Microsoft.AI.Foundry.Local; using OpenAI; using OpenAI.Chat; using System.ClientModel; using System.Diagnostics.Metrics; var alias = "phi-3.5-mini"; var manager = await FoundryLocalManager.StartModelAsync(aliasOrModelId: alias); var model = await manager.GetModelInfoAsync(aliasOrModelId: alias); ApiKeyCredential key = new ApiKeyCredential(manager.ApiKey); OpenAIClient client = new OpenAIClient(key, new OpenAIClientOptions { Endpoint = manager.Endpoint }); var chatClient = client.GetChatClient(model?.ModelId); var completionUpdates = chatClient.CompleteChatStreaming("Why is the sky blue'"); Console.Write($"[ASSISTANT]: "); foreach (var completionUpdate in completionUpdates) { if (completionUpdate.ContentUpdate.Count > 0) { Console.Write(completionUpdate.ContentUpdate[0].Text); } }

Run the code using the following command:

dotnet run

Final thoughts

The Foundry Local C# SDK empowers developers to build intelligent, privacy-preserving applications that run anywhere. Whether you're working on desktop, mobile, or embedded systems, this SDK offers a robust and flexible way to bring AI closer to your users.

Ready to get started? Dive into the official documentation:

• Getting started guide
• C# Reference documentation

You can also make contributions to the C# SDK by creating a PR on GitHub:

• Foundry Local on GitHub 

Windows 365 Cloud Apps are now available to all customers during public preview. Thank you to the customers, partners, and MVP’s who provided invaluable feedback during the private preview. Your feedback helps guide and prioritize the Windows 365 Cloud Apps feature roadmap.

What are Windows 365 Cloud Apps?

Windows 365 Cloud Apps allow administrators to give users access to specific apps delivered from the cloud instead of a full Cloud PC. This is ideal for organizations that want to streamline app delivery, reduce overhead, and modernize their virtual desktop infrastructure (VDI) environments.

Windows 365 Cloud Apps runs on Windows 365 Frontline Cloud PCs in shared mode. The Windows 365 Frontline licensing model enables shared Cloud PC access for shift-based or part-time workers, allowing unlimited users per license with one active session at a time. Since Windows 365 Cloud Apps streams only essential applications like Outlook or Word without loading a full desktop, it is ideal for task-based roles.

Together, Frontline and Cloud Apps deliver a flexible, resource-optimized solution for dynamic workforces, especially in frontline environments like retail, healthcare, and government.

New for public preview

Cloud Apps were previously supported in the Windows App during private preview. Recent changes have improved the experience for end-users, including automatically launching OneDrive and a Windows 365 filter on the Apps page of the Windows App.

Now, organizations can take advantage of a secure, scalable solution to modernize how applications are accessed and managed across their workforce. By streaming only the apps users need—without provisioning full desktop environments—Windows 365 Cloud Apps helps reduce complexity and supports flexible workstyles.

Windows 365 Cloud Apps in Windows App

Looking ahead

As organizations embrace Windows 365 Cloud Apps, Microsoft is focused on simplifying app delivery even further—making it faster, easier, and more intuitive for IT admins to deploy custom line-of-business apps at scale.

Currently, customers must create custom images to deliver custom line-of-business apps as Cloud Apps. While this works, private preview participants complained that the process is outdated.

As requested by customers, we are working to make Intune the single pane of glass for app deployment, enabling them to publish Intune Apps as Cloud Apps. Customers are already using Intune Autopilot in place of custom images, so we will similarly support Intune’s approach to modern app delivery for Windows 365 Cloud Apps. Customers will be able to include custom apps in a Cloud App policy’s Autopilot device preparation to publish them as Cloud Apps.

How to Get Started

To use Windows 365 Cloud Apps with a Windows 365 Frontline license, check out Cloud App documentation and follow the guidelines to try out the feature and evaluate Cloud Apps as an option for migrating apps to the Windows Cloud.

---

Continue the conversation. Find best practices. Bookmark the Windows Tech Community, then follow us @MSWindowsITPro on X and on LinkedIn. Looking for support? Visit Windows on Microsoft Q&A.