Seriously considering building an agent that can bulk clean up LinkedIn connection invitations using puppeteer. This has gotten out of hand

Currently finding myself checking this site multiple times a day, like maybe the number will go up at some point? www.countdown-to-reinvent.com

yeah we're into LLMs (Large Lamb-bridge Models)

TIL I learned my doctor's office guest WiFi blocks npmjs.org

Hear @clare.dev, Senior Principal Software Engineer @awscloud.bsky.social & @kateholterhoff.com discuss @kiro.dev for this RedMonk New Builders conversation. They chat about Kiro's Vibe and Spec modes, the importance of requirements, as well as Kiro's roadmap. redmonk.com/videos/tnb-c...

Kiro is my new favorite editor. It’s fantastic. I like it very much. Check it out.

Fun conversation with @jkxosound.com about where I think things are going in the AI agents space, Strands Agents, and rats 🐀

Very tired of Google search thinking I mean "Minecraft Mod Coder Pack" when I search for something related to MCP servers... 🙄

Prototyping agents is easy, fast, and fun (especially if you use Strands). But how to you get them into production with all the scalability, security, and resilience you need? We've been building Bedrock AgentCore to answer exactly that question. Check it out: aws.amazon.com/blogs/aws/in...

I sat down with Darko Mesaros to talk about Kiro, specification-driven development, and putting AI-powered development in context of the history of computing. Is an LLM a compiler? What is a spec? How are specs like SQL? Check it out: www.youtube.com/watch?v=OdD9...

Thrilled for the launch of @kiro.dev today! We started with two main ideas that led to Kiro's spec-driven development features: 1) AI can help us build better products through rapid prototyping 2) Devs can declare their app's requirements to get better results from AI, close to production-grade code

Kiro: A new agentic IDE https://kiro.dev/blog/introducing-kiro/ https://news.ycombinator.com/item?id=44560662

AWS previews Kiro IDE for developers who are over vibe coding

This is Kiro - the AI IDE that works on your messy, real-world projects. Other AI tools lose context when projects get complex. Kiro gives you spec-driven development that scales beyond prototypes. Preview: kiro.dev/blog/introducing-kiro/?trk=6b579e54-2920-4e24-a25b-d3c887a6ffdf&sc_channel=sm

Really fun conversation on the AWS She Builds series today, all things DevOps and AI agents www.youtube.com/watch?v=EawD...

Amazon targets vibe-coding chaos with new ‘Kiro’ AI software development tool https://www.geekwire.com/2025/amazon-targets-vibe-coding-chaos-with-new-kiro-ai-software-development-tool/

Why aren't there more config-based MCP servers? I see lots of MCP server code generators, but not enough that let you just provide config at runtime. I've been really enjoying: * @ivotoby/openapi-mcp-server (OpenAPI spec file -> MCP tools at runtime) * mcpdoc (llms.txt endpoints as MCP tools)

Prep for the prod launch of the first component of a major rewrite of our loan servicing app is hitting high gear. An appropriate time to sport one of my (many) favorite @charity.wtf lines.

TIL Pac-Man is in Unicode now. www.unicode.org/charts/PDF/U...

if it takes a long time to do, can you really call it "prompt" engineering

awww its little memory usage brontosauruses

If I could just get through ~one week~ without a product manager saying to me "We should build a marketplace"...

If you're at the AI Engineer World's Fair this week, there's a session on Thursday about Strands Agents

Strands Agents hit 1k 🌟 on GitHub today! Thank you to all our new contributors over the last 2 weeks, love the growing community github.com/strands-agen...

"Building an Intelligent AWS Cloud Engineer Agent with Strands Agents SDK" by Vivek V #strands-agents #agentic-ai #amazon-bedrock #aws-cdk #mcp

AWS released the Strands Agents with a Python SDK today. I spent some time tonight testing it out and deploying it on API Gateway and Lambda. I documented that exploration and hope it helps folks as they think about deploying agents on AWS. www.linkedin.com/posts/heekip...

Fun chat about Strands Agents, I love it when the space station is involved in demos www.youtube.com/watch?v=Ausm...

Excited to open source Strands Agents today! LLMs have gotten so good at reasoning and tool use that building model-driven agents with Strands is easy AND powerful Models + Tools = 🧬 Read more: aws.amazon.com/blogs/openso...

AWS today launched a new SDK for building AI agents, with support for LLMs from its own Bedrock service, LiteLLM and Ollama. By @fredericl.bsky.social

📰 New article by Clare Liguori Introducing Strands Agents, an Open Source AI Agents SDK #AWS #OpenSource

My little demo PDS has integration tests in its CI/CD pipeline now 😊

"Atomic Bloom Filters" goes so hard both as like, a great way to make super high throughput lock-free bloom filters and just as a really metal name for a programming concept

My new approach to ✨vibe-naming✨ new projects: hook together the Q Developer CLI with MCP servers for: - Thesaurus - Brave web search - GitHub - Domain name lookup And yes, I had to write a couple MCP servers to do this 😅

New AI coding benchmark: SWE-PolyBench 📊 The popular SWE-Bench evaluates on Python coding ability, only one of the most-used programming languages in the real world (StackOverflow survey chart below). SWE-PolyBench evaluates on Typescript, Javascript, Java, and Python arxiv.org/abs/2504.08703

Muhammad Shihab Rashid, et al.: SWE-PolyBench: A multi-language benchmark for repository level evaluation of coding agents https://arxiv.org/abs/2504.08703 https://arxiv.org/pdf/2504.08703 https://arxiv.org/html/2504.08703

Today we released SWE-PolyBench, a multilingual benchmark for AI coding agents. The scientists who developed the benchmark blogged about it aws.amazon.com/blogs/devops...

In a horrifying abuse of modern computers, here's a shell one-liner you can run on macOS to calculate the number of days between two dates in the most complex way imaginable - using uv and Deno and the Pydantic AI Python sandbox MCP server simonwillison.net/2025/Apr/18/...

The vast majority of MCP servers today aren't actually servers 😱 They run as a child process and communicate over stdio, a bit awkward for distributed systems trying to use MCP. Backend apps can now run existing stdio MCP servers in Lambda functions with this new library: github.com/awslabs/run-...

INVENTOR OF THE LETTER Q: yea but we're gonna make this one hella codependent

Big launch today: the Q Developer CLI now creates a website from scratch for me in under 3 minutes 🤩 aws.amazon.com/blogs/devops... (Video sped up to get under the 60 sec Bluesky limit 😊)

She's so cute

Just like we containerized webapps, we'll have to containerize AI agents... but how? What first principles led to containers in the first place, and how do they transpose to LLMs? This was my first talk at an AI meetup, but I doubt it will be the last! youtu.be/XWO_3My2eVU?...

I set up a Bluesky PDS for my little demo app www.reinvent-trivia.com - what should I do with it? Any good demo ideas?

Amazon Q Paella.

💌 AWS Valentines Thread 💌 (1/?)

I updated my serverless gen AI prompt chaining guide with examples for Bedrock Flows, alongside the examples for Step Functions. All in CDK + Python github.com/aws-samples/...

The AWS setup to run your own PDS looks pretty solid - Gonna give this a shot over the next days. Thanks for open sourcing this @clare.dev #atproto #bluesky #awscdk github.com/aws-samples/...

1/ Here the authors develop a method using semantic uncertainty to detect confabulation. In short, when an LLM generates an answer multiple times and it's truthful, the answers will be the same modulo rephrasing. When it's not, each generation will have different meaning.

It works!! I'm running a self-hosted PDS on AWS Fargate Full architecture here: github.com/clareliguori... I will confess I'm still too nervous to move my main Bluesky profile over 😅

I think about this framework all the time, since Mai-Lan first presented it in 2019. It is so so good