Skip to main content

Projects Blog About Services Contact Search

Projects Blog About Services Contact Search

Home
Blog
Tags
team

Tag: team

1 post

All adversarial ai ai-safety apple-silicon astro automation books claude cloudflare co-parenting craft economics engineering enterprise health history home-assistant homelab jailbreaking legal llm mcp meta mlx multi-agent music neurodivergence open-source organisations parenting perth philosophy policy python radio raspberry-pi research risk robotics security spark team tts tutorial typescript voice-cloning web writing

The Failure First Team

Fifteen specialist AI agents, one methodology. How adversarial AI evaluation scales through Claude Code sessions with distinct roles and standing instructions.

Site

Projects Blog Gallery Audio What's New Activity Services Search

Meta

About Now Contact Colophon Analytics

Elsewhere

GitHub LinkedIn Facebook Instagram X RSS Podcast

The methodology is public. The work is in progress. Do it properly or don't bother.

© 2026 Adrian Wedd

This site uses cookies for analytics (Google Analytics, LinkedIn Insight) and personalisation (visit count, stored locally). No tracking before consent. Full privacy details

Analytics: Google Analytics 4 tracks aggregate traffic—what gets read, where readers come from. No individual session reconstruction. Personalisation: Visit counter stored in your browser's localStorage. Never leaves your device.

Analytics Personalisation

Keyboard shortcuts

Navigation

g hHome g bBlog g pProjects g gGallery g nNow g aAbout g vActivity

Actions

/Search tToggle theme ?This help