WeAreDevelopers LIVE - Build a multi AI agents game master with Strands & our weekly web finds

Can specialized AI agents collaborate to run a D&D game? Learn how to build a multi-agent AI assistant that handles rules, characters, and complex scenarios.

#1about 1 minute

Announcing the call for papers for upcoming developer conferences

Learn about the call for papers for the Berlin World Congress and an event in San Jose, including tips for a successful submission.

#2about 9 minutes

Why Dungeons & Dragons is a great use case for AI agents

An AI-powered game master can assist human players by creating characters, managing rules, and adapting to unexpected player choices in real-time.

#3about 4 minutes

Why dedicated mobile applications for coding have failed

Mobile coding apps struggle to gain traction because developers prefer larger screens, physical keyboards, and complex multi-window development environments.

#4about 4 minutes

Examining the viability of Microsoft's AI marketplace

Microsoft's plan for an AI marketplace faces challenges, as historical attempts at third-party software marketplaces have rarely succeeded outside of native OS app stores.

#5about 6 minutes

Taking action on NPM supply chain security vulnerabilities

Following significant phishing attacks and compromises, CISA and GitHub are implementing measures to improve the security of the NPM package ecosystem.

#6about 3 minutes

Debating the necessity of retraining staff for the AI era

Accenture's decision to exit staff who "cannot be retrained" for AI raises questions about what skills are truly necessary and how to measure AI proficiency.

#7about 10 minutes

Exploring creative web projects and the rise of cozy gaming

A look at fun web finds, including IKEA-style instructions for algorithms, a "drunk mode" website theme, and the trend of "cozy gaming" for mental well-being.

#8about 7 minutes

Highlighting useful developer tools and performance insights

Discover practical developer resources like a CLI business card generator, a token pricing comparison tool, and a JavaScript performance tip using `findLast`.

#9about 6 minutes

Understanding AI agents, MCP, and the Strands SDK

Learn the fundamentals of AI agents as LLMs with tools, the Model Context Protocol (MCP) for tool access, and the A2A protocol for agent collaboration.

#10about 5 minutes

Building a simple agent and connecting to local models

A live coding demonstration shows how to instantiate a basic agent with the Strands SDK and switch from a cloud-based model to a local one using Ollama.

#11about 9 minutes

Equipping an AI agent with built-in and custom tools

Extend an agent's capabilities by providing it with built-in tools for tasks like file writing and creating a custom `roll_dice` function using the `@tool` decorator.

#12about 5 minutes

Decoupling tools by serving them on an MCP server

Learn how to run a custom tool on a standalone Model Context Protocol (MCP) server, allowing any connected agent to discover and use its functionality.

#13about 12 minutes

Creating a multi-agent system using the A2A protocol

This demonstration shows how to create a specialized character-management agent and serve it via the Agent-to-Agent (A2A) protocol for collaboration.

#14about 8 minutes

Demonstrating a complete multi-agent AI game master

See the final application where an orchestrator agent coordinates with specialized rule, character, and dice-rolling agents to run a Dungeons & Dragons game.

#15about 3 minutes

Envisioning the future of AI with agent stores

The future of personal assistants may involve an "agent store" where users connect their primary AI to specialized third-party agents to complete complex tasks.