Apr 15, 2026

A Stack Overflow for Agents? - Peter Wilson

Peter Wilson warns that AI coding assistants will continually hallucinate outdated dependencies without persistent memory. Learn how Mozilla's CQ creates a secure, localized Stack Overflow for your team.

Pause
Mute Enter Fullscreen
#1 about 3 min

The challenge of outdated training data in coding agents

How the rapid evolution of frameworks and stale model datasets cause repeated frustrations for developers.

#2 about 3 min

Introducing a shared knowledge layer for AI assistants

Enabling coding assistants to share resolutions for outdated API calls and prevent repetitive errors.

#3 about 2 min

Managing confidence scores and knowledge decay over time

Strategies for generalizing problem resolutions and managing the relevance of saved rules as systems evolve.

#4 about 4 min

Demonstrating outdated version selection in agent workflows

A live example showing an AI assistant incorrectly selecting a deprecated checkout action for a repository pipeline.

#5 about 3 min

Proposing and storing human corrections in local databases

How the system captures a human correction and saves it to a local SQLite instance for future reference.

#6 about 3 min

Applying saved knowledge units to fresh agent sessions

Launching a new prompt to verify the assistant retrieves and applies the newly corrected version rule.

#7 about 2 min

Distinguishing between local team servers and public sharing

Running the knowledge base locally versus utilizing a shared server without leaking sensitive repository data.

#8 about 4 min

Building a public knowledge commons without data scraping

Creating a decentralized, organic source of updated practices to replace stagnating technical Q&A forums.

#9 about 2 min

Mitigating malicious dependencies and knowledge poisoning in datasets

Implementing human-in-the-loop reviews and strict guardrails to prevent harmful code from entering the collective dataset.

#10 about 3 min

Gathering early community feedback and iterative deployment steps

How developers are testing the tool locally and contributing improvements through open-source pull requests.

#11 about 3 min

Applying shared knowledge to web performance and dependencies

Using agent knowledge bases to avoid bloated libraries and embrace modern built-in platform features.

#12 about 3 min

Navigating user backlash against AI in open-source products

The necessity of embracing artificial intelligence openly despite community resistance to browser integrations.

#13 about 7 min

Balancing AI enthusiasm with cynical engineering tool practices

Why developers must treat automated coding tools with a healthy skepticism instead of blind trust.

Matching moments

2:03 min

Navigating the diverse ecosystem of AI agent harnesses

Emmanuel Acheampong Emmanuel Acheampong · World Congress 2026 North America

4:06 min

Real-world case study of AI agents causing quiet instability

Vera Slavnić Vera Slavnić · Europe 2026 Virtual

2:34 min

Balancing developer autonomy with the adoption of coding agents

Clemens Wasner Clemens Wasner +4 · World Congress 2026 Europe

2:49 min

The decline of shared public knowledge in programming

Simon Auer Simon Auer · World Congress 2026 North America

1:47 min

Building a continuous learning community for AI agents

Prashanth Chandrasekar Prashanth Chandrasekar +1 · World Congress 2026 North America

4:49 min

Capturing institutional knowledge using self-learning agent traces

Mark Cavage Mark Cavage +2 · World Congress 2026 North America

Upcoming sessions on this topic

Open session

Teaching AI Coding Agents to Build It Right the First Time

  • Sajeetharan Sinnathurai

    Microsoft

    Principal Product Manager

Open session

Beware of Strangers Bearing Code: Open Source Trust in the Agent Era

  • Vikram Vaswani

    Consultant

Open session

Build a Zero-Cost AI Agent in Your Browser

  • Aishwarya Mathuria

    Adobe

    Computer Scientist

  • Nishant Thakur

    Adobe

    Senior Computer Scientist

Open session

Beyond File Dumps: Context Engineering for Coding Agents

  • Animesh Dutta

    Arm

    Senior Software Engineer

Open session

Autopsy of an Autonomous Incident: When the Agent Made It Worse

  • Navin Pai

    StackGen

    Director of Engineering

Open session

Architecting Trusted AI: Coupling Multi-Agent Systems with dbt Semantic Store for Zero-Hallucination

  • Pushkar Mishra

    J.P. Morgan

    Senior Vice President