Mike Chambers+

Something NEW! AI Functions

Strands AI Functions is a Python library for building reliable AI-powered applications through a new abstraction: functions that behave like standard Python functions, but are evaluated by reasoning AI Agents. Strands Labs Org: https://github.com/strands-labs AI Functions Code: https://github.com/strands-labs/ai-functions Blog: https://blog.mikegchambers.com/posts/software-31-ai-functions/ Mike: https://linkedin.com/in/mikegchambers

Do async tool work???

I have built some code to get true async tool calling working in an AI agent. Blog run through: https://blog.mikegchambers.com/posts/async-agentic-tools/ Code: https://github.com/mikegc-aws/async-agentic-tools Connect with me: https://linkedin.com/in/mikegchambers #coding #programming #agents #ai I work for AWS, who made Strands agents featured in this video.

Building DynamoDB Agents: Real-Time Data with Tools and Hooks

Build agents that work with Amazon DynamoDB to get access to real time data. Build tools and connect to hooks with Strands Agents SDK. Connect with me: https://linkedin.com/in/mikegchambers - Code for this video: https://github.com/mikegc-aws/strands-dynamodb-integration - Amazon Bedrock AgentCore Memory video: https://www.youtube.com/watch?v=o5A5hv82ZXU

Discussion on the potential evolution of AI agents

🎙️ Welcome to the_context(), a weekly show about Generative AI where we discuss latest news and show technical demos! In this episode, we explore what AI agents are, how they differ from traditional chatbots, and how they're evolving. We'll also introduce MCP (Model Context Protocol) as a way to decouple tools from LLMs, and highlight the growing need for multi-agent systems to handle complexity in modern AI applications. 0:00 – Intro to AI Agents 0:46 – What Is an Agent? 2:50 – Agents vs. Chatbots 4:17 – Tool Use and Loops 12:14 – What Is MCP? 16:15 – Multi-Agent Systems 21:36 – Wrap-Up

3 announcements on Amazon Bedrock that will change your prompt management

🎙️ Welcome to the_context(), a weekly show about Generative AI where we discuss latest news and show technical demos! In this episode, we introduce three new features now generally available on Amazon Bedrock: - Prompt Optimization helps tailor prompts to specific models like Claude or Titan to improve performance or allow model switching. - Prompt Routing intelligently chooses the best model variant (e.g. Claude Haiku vs Sonnet) based on prompt complexity and cost/latency trade-offs. - Prompt Caching enables faster, cheaper inference for repeated conversational patterns by reusing parts

MCP - Can Lambda do it? - Streamable HTTP Model Context Protocol

🎙️ Welcome to the_context(), a weekly show about Generative AI where we discuss latest news and show technical demos! In this episode, we deep dive into the Model Context Protocol (MCP) with lambda functions. Mike walks us through how MCP works and how you can create MCP servers with just a few lines of code—opening the door to simpler, more modular, and more scalable AI agent integrations. ---- 👥 Speakers: - Mike Chambers | Developer Advocate @AWS - Tiffany Souterre | Developer Advocate @AWS 💻 Resources: - MCP Guide: https://bit.ly/mcp-intro - MCP SDK: https://bit.ly/mcp-ts-sdk - Lam

More human-like conversations with the new Amazon Nova Sonic

🎙️ Welcome to the_context(), a weekly show about Generative AI where we discuss latest news and show technical demos! In this episode, we dive into Amazon Nova Sonic, the new speech-to-speech model from the Nova family. Unlike traditional systems, Sonic supports real-time, bidirectional communication—allowing for fluid, natural conversations and even interruptions mid-response. ----- Speakers: - Danilo Poccia | Developer Advocate @AWS - Mike Chambers | Developer Advocate @AWS - Dennis Traub | Developer Advocate @AWS - Tiffany Souterre | Developer Advocate @AWS ----- 💻 Resources: - Amazon

NEW Claude 3.7 Sonnet - Extended thinking mode in Amazon Bedrock (with code!)

Walk through how to use Claude 3.7 Sonnet extended thinking mode in Amazon Bedrock. Code: https://gist.github.com/mikegc-aws/61bd2a5b6f40ffe78f1f302aae33f620 AWS News Blog: https://aws.amazon.com/blogs/aws/anthropics-claude-3-7-sonnet-the-first-hybrid-reasoning-model-is-now-available-in-amazon-bedrock/ Anthropic News Blog: https://www.anthropic.com/news/claude-3-7-sonnet LinkedIn: https://linkedin.com/in/mikegchambers Amazon Bedrock: https://aws.amazon.com/bedrock/ [0:00] Intro [1:17] What is it? [1:55] Amazon Bedrock [3:05] The code! [5:49] Enable the thinking

Hugging Face's new smolagents with Amazon Bedrock: A Quick Guide

In this video, I demonstrate how to integrate Hugging Face's newly released Small Agents library with Amazon Bedrock, specifically using the Nova model. The video covers how to modify the default Hugging Face example to work with other model providers through Light LLM, a versatile library that enables connections to over 100 different LLM providers. I show how to convert a simple three-line Small Agents code example to work with Amazon Bedrock, and explain the unique approach of Small Agents' code-first execution strategy. Additionally, I provide a bonus demonstration of how Light LLM can be

Will Anthropic's MCP work with other LLMs? - YES, with Amazon Bedrock.

Anthropic's Model Context Protocol (MCP) is a new standard for connecting AI assistants to data sources. In this video I use Amazon Bedrock's Converse API to demonstrate how to create an MCP client with different LLMs. Code used in video is here: https://github.com/mikegc-aws/amazon-bedrock-mcp MCP Launch Post here: https://www.anthropic.com/news/model-context-protocol MCP Quickstart here: https://modelcontextprotocol.io/quickstart Connect with me on LinkedIn: https://linkedin.com/in/mikegchambers #Claude #Mistral #Meta #Llama

NEW - Amazon Bedrock's INLINE agent API

Dynamically create an agent and invoke it all in one API call. InvokeInlineAgent just released, here is a quick demo to get you started. Code from this video: https://gist.github.com/mikegc-aws/07af103cf13dea1717af645873db629c Announcement: https://aws.amazon.com/about-aws/whats-new/2024/11/inlineagents-agents-amazon-bedrock/ Boto3 for invoke_inline_agent: https://boto3.amazonaws.com/v1/documentation/api/latest/reference/services/bedrock-agent-runtime/client/invoke_inline_agent.html Previous code to create an agent: https://github.com/build-on-aws/agents-for-amazon-bedrock-sample-feature

Claude Rickrolled me! - Sonnet 3.5 v2 Computer Control

Don't do this! My code is here: https://github.com/mikegc-aws/claude-computer-control Anthropic's code is here: https://github.com/anthropics/anthropic-quickstarts/tree/main/computer-use-demo

I'm Learning | Can I automate the build? | Amazon Bedrock Knowledge Base

This is me learning in the open! In this video I talk about how I am tackling the fully automated creation of an Amazon Bedrock Knowledge Base (managed RAG), from scratch, running on an OpenSearch Serverless collection. Come and connect https://linkedin.com/in/mikegchambers

Where do you want to go? The future please!

I was lucky enough to experience full self driving on a recent trip to in San Francisco. I am all about this... wake me up when it's everywhere.

I used it! - Claude 3 Haiku on Amazon Bedrock - NodeJS text and image prompt.

YES! 🎉 Claude 3 Haiku! - 👨‍💻 Let's DIVE DEEP and WRITE SOME CODE using Amazon Bedrock. 👉 Setup access to Claude 3 Haiku in Amazon Bedrock 👉 Use the Amazon Bedrock playground to experiment with text and images. 👉 Write some NodeJS code* that sends a text prompt to Haiku. 👉 Write some more NodeJS code* that sends text AND an image to Haiku. Links: - Code Samples: https://lnkd.in/ggGS2H54 - Community.aws: https://lnkd.in/ghspNEHp - Code from this post: https://lnkd.in/gUiFMfmz (yes I ended up putting it on Github 😁 ). What do you think of this post? Just right? Wrong orientation? Let m