Good morning, everyone!
The big news this week is from Ring, which announced Ring Robotics with two new devices: the Ring Drone Cam, an indoor flying security camera for the home, and the Ring Rover Cam, a roving camera that patrols a business. Check out the video demo of the Ring Rover and here's the Drone Cam:
AWS launched AWS Emulator, a new local emulator of AWS services for testing and CI, and shared how they built the agents that wrote it. Over on All Things Distributed, Swami Sivasubramanian looks at the engineer's job when agents write the code.
The Kiro team released Kiro for Android and introduced "Approve for me", a new way to auto-approve. If you've been wondering about Kiro Crew, there's a getting started video and a walkthrough of what Kiro Crew is. It's also Kiroween right now!
- 🏠 Alexa+ Smart Home: Matter updates — Matter-enabled Echo devices now support Matter 1.5.1, and Echo Show 16, Echo Time, and Echo Studio with Home Camera now work as Matter controllers.
- 📷 Ring Appstore: Watchtower sample app v1.1.0 — Our open-source Watchtower sample covers account linking, verified webhooks, and multi-camera live video, and the release notes add a certification preflight checklist.
- 🔑 Amazon Data Portability: scope reference — New docs map every scope a customer can authorize your app to receive, with schemas and sample output for Category-1, Category-2, and Category-3 data.
- 🛠️ AWS SDK for .NET 4.1 — AWS SDK for .NET 4.1 is out, the first release on its annual minor-release cycle, with reworked dependencies for better performance and simpler setup.
- 👻 Kiro: "Approve for me" — Kiro IDE, CLI, and Web add "Approve for me," a model-based review that auto-approves routine agent actions, blocks risky ones, and asks about the rest.
🤖 Ring Robotics launch
- Ring just announced Ring Robotics, launching with the indoor Ring Drone Cam for homes and the patrolling Ring Rover Cam for businesses, both running on its AI Pilot mapping system.
🧪 AWS Emulator launch
- Launched this week, AWS Emulator runs core services like S3, DynamoDB and SQS locally on a laptop or in a CI job, so the app or skill backend code that calls AWS can be tested without a deploy or an AWS account.
👻 Kiro for Android
- If you run Kiro cloud agents, the new Kiro for Android app on Google Play lets you start, steer, and review agent sessions from your phone and approve their diffs as pull requests.
This season we're heading to AWS re:Invent 2026 in Las Vegas, where my colleague Kourtney is leading the workshop "Turn a Bee wearable into your re:Invent notetaking companion," and we hope to see you there! If you're building from home, Kiroween 2026 is an online hackathon worth knowing about.
| Date | Event | Location | Description |
|---|---|---|---|
| Oct 30 – Dec 4 | Kiroween 2026 | Online | Kiro's annual Halloween-themed hackathon: build a new app with Kiro in one of four categories for a share of $100,000+ in prizes, with submissions due December 4. |
| Nov 30 – Dec 4 | AWS re:Invent 2026 | Las Vegas, US | AWS's flagship annual conference, with 1,000+ technical sessions, keynotes, workshops, and the largest cloud-community gathering of the year. Dev Advocate Kourtney Meiss is leading a workshop, "Turn a Bee wearable into your re:Invent notetaking companion." |
AWS Emulator gets two inside looks this week, one on how autonomous agents built it and one from Swami Sivasubramanian on what an engineer's job becomes when agents write the code. There's also a practical guide to giving those agents AWS access safely, and our latest DEV Chat on what the React Foundation means for React Native developers.
📽️ What the React Foundation Means for React Native Devs | DEV Chat
React now has a home outside Meta, and in this DEV Chat episode recorded at NextApp DevCon in Berlin, Seth Webster, Head and Founder of the React Foundation, tells Steven Moyes, Sr. Product Manager at Amazon, what that shift changes for the people building on it. They dig into how Vega OS pushed React Native past iOS and Android, why a stable API matters on the road to 1.0, and how to get your own fixes upstream. For Fire TV teams on React Native, it explains where the framework under your app is going and how to help steer it.
📰 How we built the agents that built AWS Emulator
Sr. Principal Engineer Matthew Goldberg tells the story behind AWS Emulator, a local emulator of AWS services whose code was written entirely by autonomous agents. No one on his team wrote or reviewed any of it. He traces how a weekend experiment with Kiro became a product, and makes the case that agents need well-defined boundaries, his "coloring books," before you can trust them to work unsupervised. The guardrails are the part worth borrowing for anyone giving agents bigger pieces of a codebase, and the emulator is a practical way to test AWS-backed app code on your own machine.
📽️ Harness Engineering: Context, Skills, and Reliable AI Workflows | React Universe On Air
Our Dev Advocate Anisha Malde sat down with Callstack's Łukasz Chludziński at Agent Conf 2026 in Warsaw to unpack harness engineering, the practice of giving coding agents the right context, tools, and feedback so their output holds up. She grounds it in Amazon Devices Builder Tools and Vega, including Fire OS-to-Vega migration and reusable agent skills. It's a useful listen for anyone handing more of their Vega or React Native work to an agent who still wants to own the user experience and the technical judgment behind it.
📽️ Automated Reasoning 101
Brooke Jamieson breaks down automated reasoning for the AWS Developers channel, explaining it as the logic-based way of proving a rule holds for every case instead of testing a sample of them. A single fictional online shop runs through the whole explainer, and the same question gets asked of each part of it: what must never happen? She also shows where AWS already applies these proofs to access policies and network paths, and why a correct proof of the wrong rule gives you false confidence. It's a good primer for anyone building AI agents or handling customer data, because it shows what a formal guarantee covers and where it stops.
📰 AWS Emulator and the engineer's job when agents write the code
On All Things Distributed, Swami Sivasubramanian marks the launch of AWS Emulator, a local stand-in for more than 200 AWS services that agents wrote and tested end to end, by recalling the high-fidelity simulation his team built to test one of Amazon's earliest distributed systems. The larger thread is how the engineer's job changes once agents write the code, from someone who has been testing distributed systems at Amazon since the early days. It's a useful read for anyone who tests a cloud backend before it ships, and for a grounded view of where engineering judgment goes next.
📰 A Practical Guide to AWS Permissions for AI Coding Agents
On the AWS Developers channel, Saurabh Rob Dahal starts from the story of an AI coding agent that deleted a production database after being told not to. His point is that the agent's credentials allowed the delete, so the fix is in how you scope its IAM access, not in what you put in the prompt. The video shows how to give agents like Claude Code, Kiro, or Codex only the access a task needs and keep production out of reach. A companion AWS Builder Center guide to permissions for AI coding agents covers how to replace leftover admin access with a narrowly scoped role that uses temporary credentials, and how to have the agent draft those policies for you to review.
📽️ Debugging 101 for Vibe Coders
A short explainer on the AWS Developers channel from Fabricio Flores takes on the moment every vibe coder knows: the agent keeps guessing, and "try again" only makes it worse. Fabricio argues that the agent gets stuck because you've given it nothing new to work with, and lays out the habits that break the loop: describing the bug precisely, giving the agent the right context, and protecting your work before it starts editing. Five minutes here will save you a lot of reruns the next time a fix turns into a fresh bug.
📰 How Amazon Leo satellite internet works
By Alex Davies
Alex Davies breaks down how Amazon Leo's growing constellation of custom-built satellites in low Earth orbit works together to deliver internet connectivity anywhere on the planet. It's a clear primer on the system as a whole, and the fastest way to build a working mental model of the network without wading through technical filings.
📽️ What Kiro Crew Is and When to Use It vs IDE, CLI, and Web
On the Kiro channel, Giovana Armani introduces Kiro Crew in a short explainer. Kiro Crew is the new interface that coordinates several agents at once, and she places it next to the Kiro IDE, CLI, and web versions. All four run on the same Kiro AI Assistant underneath, so the choice comes down to how you want to work and which features each surface gives you. Watch it to see which one matches your workflow before you set any of them up.
📽️ How to Get Started with Kiro Crew
Kiro Crew runs the Kiro agent as a service on your own machine so work continues after you step away, and Navyata Bawa demos it on the Kiro channel. The tutorial builds a repo watcher that remembers how you work, investigates issues in parallel, fixes bugs test first, and pings you only when something needs a decision. It's a hands-on way to see what a long-running agent can take off your plate before you hand it your own repos.
📰 Can we ship a secure product with vibe coding?
By Purnaresa Y
Purnaresa Y, a Senior Security Solutions Architect at AWS, uses recent benchmarks and real CVE data to ask whether a vibe-coded first draft is safe to ship. The piece sorts the risk into three problems: output that changes from run to run, working code that still has vulnerabilities, and agents that take actions nobody asked for. Each one gets a fix that lives in how you build with the agent, not in a cleverer prompt. It's a useful checklist for Amazon developers who let an AI agent write production code for their Fire TV, Alexa or Appstore apps.
📽️ Amazon AWS CEO: AI boom is just getting started
AWS CEO Matt Garman sits down with Alex Heath on the Sources podcast to explain why he believes AI inference is still in its earliest days, and what that means for the cloud business AWS spent two decades building. The conversation ranges from how AWS decides who gets scarce GPU capacity to the reset inside the Nova team and the push to sell Trainium chips outside AWS. For anyone shipping AI features, it's a candid look at where the compute underneath them is headed and how tight it's likely to stay for the next couple of years.
📬 Feedback
Questions or comments? Have an item for next week's post? Let me know!




Top comments (0)