Feeds

A few months ago, we introduced the first preview releases of the AI and Vector Data extensions—powerful .NET libraries designed to simplify working with AI models and vector stores. Since then, we’ve been collaborating closely with partners and the community to refine these libraries, stabilize the APIs, and incorporate valuable feedback. Today, we’re excited to announce that these extensions are now generally available, providing developers with a robust foundation to build scalable, maintainable, and interoperable AI-powered applications. ...

The Internet relies on the Border Gateway Protocol (BGP) to exchange IP address reachability information. This information outlines the path a sender or router can use to reach a specific destination. These paths, conveyed in BGP messages, are sequences of Autonomous System Numbers (ASNs), with each ASN representing an organization that operates its own segment of Internet infrastructure. Throughout this blog post, we'll use the terms "BGP routes" or simply "routes" to refer to these paths. In essence, BGP functions by enabling autonomous systems to exchange routes to IP address blocks (“IP prefixes”), allowing different entities across the Internet to construct their routing tables. ...

It’s been a great day at Google I/O, and I think we made a big industry impact. I still have to catch up on what we wrote, and what others wrote about us, so expect a few of those in tomorrow’s reading list. [blog] What’s new with Agents: ADK, Agent Engine, and A2A Enhancements. Big updates for agent developers, including a Java version of the Agent Development Kit. [blog] Announcing Gemma 3n preview: powerful, efficient, mobile-first AI. Open, small, and powerful. This looks like a great option for mobile or local workloads. ...

Feels like Friday. Only Wednesday. I’m home after Google I/O and a customer meeting, and another customer chat tomorrow. I’m learning a lot! [blog] Understand all the I/O news with NotebookLM. This is a creative and smart way to make it simpler to digest and explore the news from Google I/O. [blog] Google I/O 2025: From research to reality. Take a moment to read Sundar’s post and get a good idea of where AI is now, and heading. ...

Top Links How Microsoft Developers Use AI in Real-World Coding | BRK103 | Microsoft Build 2025 (David Fowler & Stephen Toub) Introducing Microsoft 365 Copilot Tuning, multi-agent orchestration, and more from Microsoft Build 2025 (Jared Spataro) Building with AI: highlights for developers at Google I/O (Mat Velloso) What’s new in Flutter 3.32 (Kevin Chisholm) Updates to the Android XR SDK: Introducing Developer Preview 2 (Matthew McCullough) Introducing the Microsoft 365 Agents Toolkit (Zhidi Shang) ...

Authors: Sejoon Oh, Moumita Bhattacharya, Yesu Feng, Sudarshan Lamkhede, Ko-Jen Hsiao, and Justin Basilico MotivationRecommender systems have become essential components of digital services across e-commerce, streaming media, and social networks [1, 2]. At Netflix, these systems drive significant product and business impact by connecting members with relevant content at the right time [3, 4]. While our recommendation foundation model (FM) has made substantial progress in understanding user preferences through large-scale learning from interaction histories (please refer to this article about FM @ Netflix), there is an opportunity to further enhance its capabilities. By extending FM to incorporate the prediction of underlying user intents, we aim to enrich its understanding of user sessions beyond next-item prediction, thereby offering a more comprehensive and nuanced recommendation experience. ...

Where you decide to run your web app is often a late-binding choice. Once you’ve finished coding something you like and done some localhost testing, you seek out a reasonable place that gives you a public IP address. Developers have no shortage of runtime host options, including hyperscalers, rented VMs from cheap regional providers, or targeted services from the likes of Firebase, Cloudflare, Vercel, Netlify, Fly.io, and a dozen others. I’m an unapologetic fanboy of Google Cloud Run—host scale-to-zero apps, functions, and jobs that offer huge resource configurations, concurrent calls, GPUs, and durable volumes with a generous free tier and straightforward pricing—and we just took the wraps of a handful of new ways to take a pile of code and turn it into a cloud endpoint. ...

When using GitHub as the version control system (VCS) provider for Terraform, it’s hard to decide which authentication methods the organization should adopt: GitHub App or GitHub OAuth? In this blog, we’ll explore the differences between these two methods, their advantages, drawbacks, and best practices. This will help you choose the right approach based on your team’s security and operational requirements. Whether you're managing a small project or an enterprise Terraform setup, understanding these options will help you optimize your workflow and security posture. ...

Imagine this: Someone intercepts your private messages, secures your encrypted business files, or captures classified government communications —but they can’t read any of it. Not yet anyway. But they’re not trying to crack the code today. They’re just collecting. They’re waiting. This is the strategy behind harvest now, decrypt later (HNDL) attacks — a looming threat in the age of quantum computing. What is a harvest now, decrypt later attack? At its core, HNDL is about playing the long game. The idea is simple but powerful: Attackers collect encrypted data now, knowing that in the future, quantum computers may be able to break the encryption protecting it. ...