The AWS Developers Podcast

The MCP specification just went stateless. Clare Liguori — Senior Principal Engineer at AWS and core MCP maintainer — explains what changed, why it matters, and what it unlocks for agent developers building with Strands Agents and Kiro. In this episode, Romain sits down with Clare Liguori, Senior Principal Engineer at AWS, to discuss the July 28 MCP spec release, new MCP extensions (Skills, Tasks, Events), Strands Agents Harness SDK, Strands Shell, Physical AI in Strands Labs, and the latest Kiro updates across Web, CLI, IDE 1.0, and iOS. Key takeaways: • MCP goes stateless — the July 28 spec release removes the need for stateful streaming in remote MCP servers, so SaaS providers can drop sticky sessions and load-balancer gymnastics and run each request anywhere. Expect a new wave of remote MCP servers over HTTP. • New MCP extensions framework — features now start as stable extensions before graduating into the official spec: Skills over MCP (bundle a workflow and its tools together), long-running Tasks (kick off builds or jobs without blocking the agent), and Events (trigger always-on agents from external signals like Slack or an earthquake feed). • Strands Agents Harness SDK & TypeScript 1.0 — a more batteries-included harness with context management, compaction, and excellent out-of-the-box file tools, plus the TypeScript SDK reaching 1.0. Upgrade the model ID and your agent gets better. • Strands Shell — a lightweight, in-process agent sandbox written in Rust (cross-platform, Python SDK today). It gives an agent a virtual file system and minimal bash/Lua scripting without a heavyweight VM — great as a safe scratch pad or for scripting tools together. • Strands Labs & Physical AI — an experimental space for bleeding-edge agent ideas, including combining low-latency local VLA models on robots with the long-range, multi-task reasoning of frontier models in the cloud. • AgentCore Managed Harness — a configuration-based way to run agents (prompt, model, Lambda tools, context and session management) that is Strands under the hood, no Python or TypeScript required. AgentCore Gateway added MCP 2026-07-28 support day one, with version negotiation for backward compatibility. • One unified Kiro harness — Kiro Web, iOS, CLI, and IDE 1.0 now share one harness, bringing spec-driven development, hooks, skills, and powers to every client and letting new features ship across clients on the same day. • Automated reasoning in specs — property-based testing plus ambiguity and conflict detection in requirements help you express intent clearly; the permission system is built on Cedar with policy presets like dev shell, trust all, and read all. • Right-sizing specs and collaborating — check specs into code as a snapshot of intent, watch design and task-list length as a signal to split into multiple specs, add per-task validation steps, and collaborate on specs with comments in Kiro Web.

What happens when you strip the model from an agent? Everything left is the harness — and engineering that harness is the new discipline every AI builder needs to master. In this episode, Romain sits down with Mike Chambers, Senior Developer Advocate for Generative AI at AWS, fresh from speaking at the AI Engineers World's Fair 2026. They dig into harness engineering — what it means, why it matters, and how it changes whether you're building or using agents. Mike introduces “slop ops” (his term for unconstrained agent deployments), explains the difference between Strands Agents SDK and AgentCore Harness, and shares why small language models are his next big bet. Key takeaways: • Harness engineering is the discipline around everything left when you remove the model from an agent — including tools, skills, memory, context, observability, evaluations, and the agentic loop • The agent you use and the agent you build require different harness thinking — personal coding agents can optimize for productivity, while production agents must manage context, cost, scale, and reliability • AgentCore Harness removes undifferentiated orchestration code — provide a model, system prompt, tools, and skills through configuration, and the service creates and runs the agentic loop • Avoid “slop ops” — agents should generate infrastructure as code rather than directly creating unconstrained cloud resources, keeping deployments repeatable, reviewable, and owned by the team • Evaluation must evolve alongside an agent — it is difficult to retrofit and cannot be treated as a one-time prerequisite or an afterthought • FOMAT, the fear of missing agent time, can push developers toward unhealthy always-on behavior — faster execution still requires space to think deeply about which problems are worth solving • Kiro for iOS enables mobile-first agentic workflows — Mike uses walks to think, dictate ideas, delegate experiments, and return to working prototypes • Small language models can make focused agentic workloads faster and more cost-effective once teams understand the task well enough to specialize the model

Prime Video serves over 300 million monthly customers across 240+ territories and 8,000+ device types. Lilia Abaibourova leads the AI-native transformation there. Prime Video started early, adopting CodeWhisperer as soon as it became available in preview, progressing through Amazon Q Developer, and now using Kiro to put agentic development in the hands of engineers, product managers, and designers. Romain sits down with Lilia to trace that journey — from CodeWhisperer preview to Amazon Q Developer and now Kiro — and explore MCP as the organizational inflection point: teams run two-person pods with full decision autonomy, while managers oversee fleets of agents instead of just headcount. Key takeaways: • Prime Video adopted AI coding tools from day one — CodeWhisperer in preview, then Amazon Q Developer, and now Kiro — building institutional muscle even when specific features were later sunset • MCP was the inflection point for organizational (not just individual) transformation — connecting disciplines end-to-end and eliminating handoff bottlenecks • Code generation is now effectively solved; the new frontier is verification, validation, and code review at scale • Team structures are being actively experimented with — from 2-person pods with full autonomy to larger teams where managers leverage agent fleets • The biggest anti-pattern: waiting on the sidelines for a standard instead of learning through experimentation

I keep hearing 'we're building an AI platform' from enterprises. Most are stitching together demos. Thomson Reuters has been doing AI for 30 years. I sat down with Laura Skylaki (VP AI Engineering, Thomson Reuters) at AWS Summit London — and the gap between what they're doing vs. what most companies call 'AI maturity' is striking. In this episode, Laura breaks down how Thomson Reuters built an enterprise AI platform that serves 30,000 employees across three user personas — from business users to AI creators to developers — with 16 internal services, a $200M/year AI commitment, and a healthy skepticism about when agentic is actually the right answer. Key takeaways: • Evaluation before building — 'It's increasingly harder to retrofit an evaluation framework to an agent capability you already deployed.' They build evaluation benchmarks first to force understanding of the business problem, then build the agent. LLM-as-a-judge with human-in-the-loop calibration — and the judges themselves need continuous regression checks. • Know when NOT to use agents — 'Where we can go and write deterministic code with a logical rule-based approach, I don't think it makes sense to rely on the agent.' Not everything needs to be agentic. When tools exceed a dozen or overlap semantically, specialize into multiple agents instead of overloading one. • Engineers haven't written code manually in 3-4 months — Not experimenting; their AI coding tooling is that far ahead. The shift is toward system design, architecture, and clear structuring for AI assistance. • AWS under the hood with AgentCore and Strands — The platform runs on AWS with API Gateway, Lambda, Step Functions, and Amazon Bedrock AgentCore for agent deployments (runtime, memory, MCP marketplace, code interpreter, browser tools, policy and identity governance). They use Strands Agents SDK to build agents. AgentCore removes the infrastructure burden of long-running agent sessions. • Rethink workflows, don't retrofit — Teams reimagining legacy workflows from scratch for AI rather than bolting generative AI onto systems not built for it. 'If I was building this now for the first time, how would I build it?' The companies that succeed with AI are not fitting it into existing processes — they're redesigning the process through AI.

Brenn reached out via LinkedIn to share his experience with AI-driven software development and his new book, The Delivery Gap. Romain read a copy during a business trip and found it deeply aligned with how he guides customers through their AI transformation — and discovered a few new angles worth exploring, including a convergence on cost tracking that maps directly to Amazon's internal cost to serve software metric. In this episode, Brenn — Senior Manager at Delivery Hero (one of the world's largest food delivery companies, operating in 65 countries) — breaks down why most companies fail to see returns from AI coding tools despite individual developers feeling more productive. The core insight: generating code 10x faster means nothing if your verification infrastructure can't keep up. You're just driving 10x faster into a wall. Key takeaways: • The 96% waste problem — If you generate 100 PRs and only 4 make it to production and stay there, the other 96 are waste. Measuring PRs created is meaningless; measure what ships and survives. • The verification triangle — Your delivery speed is governed by verification infrastructure, not generation speed. Banks can't release faster than they can audit. Find your constraint — that's where investment should go, not more coding tools. • Cost per accepted change — Total token costs + human time for all PRs, divided by changes that reach production and stay there. This single metric reveals where waste accumulates and aligns with Amazon's cost to serve software model. • Specs as alignment documents, not source code — Specs align humans and AI on intent and why, not for deterministic code generation. The same spec produces different software each time. Focus on why; let the AI document the what. • Keep agents small and focused — Every MCP server re-injected into context is a cost multiplier per turn. The smallest, tightest, most precisely aimed agent outperforms a Swiss Army knife agent on both cost and accuracy. Apply cost per accepted action to measure agentic ROI.