OReilly/Event/26/03/AI Codecon

• AI Codecon: Software Craftsmanship in the Age of AI

  • OReilly/Event/26/03/AI Codecon/People

  • When

    • 2026-03-26 Thu
    • Start: 11:00 EDT (11am Eastern)
    • End: 15:00 EDT (3pm Eastern)
    • Window: 11:00–15:00 EDT (four hours)

  • Agenda source

    • O’Reilly’s public schedule markup lists segment order and length (10 min, 25 min, …) for each <li>—that is the official minute-by-minute agenda; it does not include separate clock start/end fields.
    • Timeline below assigns wall-clock slots in EDT by cumulating those lengths from the 11:00 EDT start; the listed segments total 3h 55m, so the programme ends 14:55 EDT with 5 minutes before the 15:00 EDT end time.

  • Timeline (11:00–15:00 EDT)

    • 11:00–11:10 — Introduction — OReilly/Event/26/03/AI Codecon/1100 Introduction Tim OReilly and Addy Osmani
    • 11:10–11:35 — Fireside — OReilly/Event/26/03/AI Codecon/1110 Fireside Chat Cat Wu and Addy Osmani
    • 11:35–11:55 — Antfarm / OpenClaw — OReilly/Event/26/03/AI Codecon/1135 Antfarm OpenClaw Team of Agents Ryan Carson
    • 11:55–12:15 — Orchestrating coding agents — OReilly/Event/26/03/AI Codecon/1155 Orchestrating Coding Agents Addy Osmani
    • 12:15–12:20 — Break (5 min)
    • 12:20–12:40 — Hidden cost of agentic failure — OReilly/Event/26/03/AI Codecon/1220 Hidden Cost of Agentic Failure Nicole Koenigstein
    • 12:40–13:05 — Prompt → multi-agent system — OReilly/Event/26/03/AI Codecon/1240 From Prompt to Multi-Agent System Hila Fox
    • 13:05–13:10 — When agents break production — OReilly/Event/26/03/AI Codecon/1305 When AI Agents Break Production Advait Patel
    • 13:10–13:15 — Is your eval lying — OReilly/Event/26/03/AI Codecon/1310 Is Your Eval Lying Abhimanyu Anand
    • 13:15–13:20 — Break (5 min)
    • 13:20–13:40 — Mythical Agent-Month — OReilly/Event/26/03/AI Codecon/1320 Mythical Agent-Month Wes McKinney
    • 13:40–13:50 — Automation → augmentation / AI coaches — OReilly/Event/26/03/AI Codecon/1340 From Automation to Augmentation Mike Amundsen
    • 13:50–13:55 — Agent provenance — OReilly/Event/26/03/AI Codecon/1350 Agent Provenance Tatiana Botskina
    • 13:55–14:00 — MCP server testing — OReilly/Event/26/03/AI Codecon/1355 Automating MCP Server Testing Neethu Elizabeth Simon
    • 14:00–14:05 — Personalized messaging at scale — OReilly/Event/26/03/AI Codecon/1400 Personalized Messaging LangGraph Arushee Garg
    • 14:05–14:10 — Break (5 min)
    • 14:10–14:30 — AI Flower — OReilly/Event/26/03/AI Codecon/1410 AI Flower Juliette van der Laarse
    • 14:30–14:50 — Fireside — OReilly/Event/26/03/AI Codecon/1430 Fireside Chat Aaron Levie and Tim OReilly
    • 14:50–14:55 — Closing — OReilly/Event/26/03/AI Codecon/1450 Closing Remarks Tim OReilly and Addy Osmani

  • Resources

    • Event page: O'Reilly AI Codecon
    • Code of Conduct: Link

  • Schedule — Links

    • OReilly/Event/26/03/AI Codecon/Pt 1 Opening and Orchestration

      • Part 1: Opening and Orchestration

        • Introduction

          • Session: OReilly/Event/26/03/AI Codecon/1100 Introduction Tim OReilly and Addy Osmani
          • Speakers: Person/Tim O'Reilly and Person/Addy Osmani
          • Length: 10 min
          • Description: Welcome to Software Craftsmanship in the Age of AI.

        • Fireside Chat

          • Session: OReilly/Event/26/03/AI Codecon/1110 Fireside Chat Cat Wu and Addy Osmani
          • Speakers: Person/Cat Wu and Person/Addy Osmani
          • Length: 25 min
          • Description: How the software engineer role is shifting in 2026; Claude Code and Anthropic’s integration of AI-assisted development (~90% of Anthropic’s code written by Claude).

        • How to Create a Team of Agents in OpenClaw and Ship Code with One Command

          • Session: OReilly/Event/26/03/AI Codecon/1135 Antfarm OpenClaw Team of Agents Ryan Carson
          • Speaker: Person/Ryan Carson
          • Length: 20 min
          • Description: Antfarm on OpenClaw—multi-agent YAML workflows (Planner, Developer, Verifier, Tester), Ralph loop, verification gates, shipping with one command.

        • Orchestrating Coding Agents

          • Session: OReilly/Event/26/03/AI Codecon/1155 Orchestrating Coding Agents Addy Osmani
          • Speaker: Person/Addy Osmani
          • Length: 20 min
          • Description: Patterns for parallel agents, isolated workspaces, decomposition, and orchestration as a first-class engineering skill.

        • Break — 5 min

    • OReilly/Event/26/03/AI Codecon/Pt 2 Production Systems and Evaluation

      • Part 2: Production Systems and Evaluation

        • The Hidden Cost of Agentic Failure and the Next Phase of Agentic AI

          • Session: OReilly/Event/26/03/AI Codecon/1220 Hidden Cost of Agentic Failure Nicole Koenigstein
          • Speaker: Person/Nicole Koenigstein
          • Length: 20 min
          • Description: Why multi-agent systems fail at the pipeline level—error propagation, handoffs, cost, and system-level optimization vs better single agents.

        • From Prompt to Multi-Agent System: The Evolution of Our AI Product

          • Session: OReilly/Event/26/03/AI Codecon/1240 From Prompt to Multi-Agent System Hila Fox
          • Speaker: Person/Hila Fox
          • Length: 25 min
          • Description: Qodo’s path from one agent to multi-agent code review with LangGraph and MCPs.

        • When AI Agents Break Production and What We Did About It

          • Session: OReilly/Event/26/03/AI Codecon/1305 When AI Agents Break Production Advait Patel
          • Speaker: Person/Advait Patel
          • Length: 5 min
          • Description: Broadcom war story—an agent change that passed tests but degraded production hours later.

        • Is Your Eval Lying to You? Catching Hidden Failures in Agent Evaluation

          • Session: OReilly/Event/26/03/AI Codecon/1310 Is Your Eval Lying Abhimanyu Anand
          • Speaker: Person/Abhimanyu Anand
          • Length: 5 min
          • Description: Hidden trade-offs, sycophantic judges, and significance testing—Elastic’s experiment design.

        • Break — 5 min

    • OReilly/Event/26/03/AI Codecon/Pt 3 Skills Provenance and Tooling

      • Part 3: Skills, Provenance, and Tooling

        • The Mythical Agent-Month

          • Session: OReilly/Event/26/03/AI Codecon/1320 Mythical Agent-Month Wes McKinney
          • Speaker: Person/Wes McKinney
          • Length: 20 min
          • Description: Parallel Claude sessions and speed vs compounding complexity—Brooks’ lessons when code is “free.”

        • From Automation to Augmentation: Designing AI Coaches That Amplify Expertise

          • Session: OReilly/Event/26/03/AI Codecon/1340 From Automation to Augmentation Mike Amundsen
          • Speaker: Person/Mike Amundsen
          • Length: 10 min
          • Description: Skill formation vs one-off answers; Engelbart/Licklider; demo of an AI coach pattern.

        • Agent-to-Agent Collaboration and Agent Provenance

          • Session: OReilly/Event/26/03/AI Codecon/1350 Agent Provenance Tatiana Botskina
          • Speaker: Person/Tatiana Botskina
          • Length: 5 min
          • Description: Agent registries, ownership, versioning, lineage, and trust for composable multi-agent systems.

        • Automating MCP Server Testing: Engineering Reliability for Agentic Systems

          • Session: OReilly/Event/26/03/AI Codecon/1355 Automating MCP Server Testing Neethu Elizabeth Simon
          • Speaker: Person/Neethu Elizabeth Simon
          • Length: 5 min
          • Description: Arm’s MCP server tested with pytest, Testcontainers, and real servers in CI—not mocking LLM outputs.

        • How Multi-Agent Systems Enable Personalized Messaging at Scale

          • Session: OReilly/Event/26/03/AI Codecon/1400 Personalized Messaging LangGraph Arushee Garg
          • Speaker: Person/Arushee Garg
          • Length: 5 min
          • Description: LinkedIn Premium’s supervisor-based LangGraph workflow for tailored jobseeker messages.

        • Break — 5 min

    • OReilly/Event/26/03/AI Codecon/Pt 4 Capability Architecture and Closing

      • Part 4: Capability Architecture and Closing

        • The AI Flower: A Public Capability Architecture for AI-Native Engineering

          • Session: OReilly/Event/26/03/AI Codecon/1410 AI Flower Juliette van der Laarse
          • Speaker: Person/Juliette van der Laarse
          • Length: 20 min
          • Description: A capability model to navigate overload—skill fossilization lens, mapping practice and gaps without lowering quality bar.

        • Fireside Chat

          • Session: OReilly/Event/26/03/AI Codecon/1430 Fireside Chat Aaron Levie and Tim OReilly
          • Speakers: Person/Aaron Levie and Person/Tim O'Reilly
          • Length: 20 min
          • Description: Box, intelligent content management, context for AI, file systems as core primitives, API-first software, agentic identities.

        • Closing Remarks

          • Session: OReilly/Event/26/03/AI Codecon/1450 Closing Remarks Tim OReilly and Addy Osmani
          • Speakers: Person/Tim O'Reilly and Person/Addy Osmani
          • Length: 5 min
          • Description: Wrap-up for the day.