Agents Don’t Ship Themselves
summary
Claims that demos won’t run themselves, backed by an early Rust agentic coding system built on Amazon Bedrock that mostly worked but felt like “sticks and tape.” Shows an open-source Strands SDK example where one line (line 18) creates an agent and tools, a two‑agent orchestrator (web fetch + calculator), and deployment to Bedrock Agent Core via CLI. Covers memory embedding delay, Agent Core Gateway, code execution and browser tooling, and which pieces required extra ops to be production ready.
AI-written summary · about
chapters (10)
- 0:00Agent deployment challenges
- 1:21Building production agents
- 3:49Strands agent framework
- 5:32Multi-agent collaboration
- 6:58Bedrock Agent Core deployment
- 8:42Live demo: cloud invocation
- 10:08Agent memory management
- 12:31Bedrock Agent Core features
- 13:54Cloud deployment necessity
- 14:14Learning resources
related talks
19mBuilding Agents at Cloud Scale
01h 21mShip Agents that Ship: A Hands-On Workshop
20mShip it! Building Production Ready Agents
33mShipping Enterprise AI Agents with Amazon Bedrock AgentCore
34mThe Hard Truth About AI Agents in Production
06mKeynote: Building Agents at Scale
10mAgentCraft: Putting the Orc in Agent Orchestration
06mSponsored Keynote: Day-2 Ready: Bringing ag...