About

I'm Adrian Wedd. I live in Tasmania and I build things that think, break, and sometimes both.

By day I work in systems integration and cybersecurity for Tasmania's public housing sector. By every other hour I'm an independent AI safety researcher—building the multi-agent systems I study, writing about what happens when demonstrated risk meets institutional optimism, and trying to make more of that work public.

My background spans environmental advocacy, government service, infrastructure management, and a persistent obsession with how complex systems fail. This site is where those threads converge.

What I'm Working On

AI safety frameworks that start from failure cases rather than success metrics. Multi-agent coordination with physical safety constraints. The structural gap between what organisations know about technological risk and what they do about it.

I also write, make art, and occasionally record audio. This site is my workshop—not a portfolio, not a CV, but a space where finished work, ongoing experiments, and raw thinking coexist.

Elsewhere

• GitHub — 115+ repos across AI, security, music tools, and more
• Failure First — AI safety methodology
• This Wasn't in the Brochure — a book about parenting and neurodivergence

Privacy

This site respects your choices. No tracking happens without consent. Analytics (if enabled) uses Google Analytics 4 with cookie flags set to SameSite=Lax and Secure. Personalisation stores visit data locally in your browser — nothing leaves your machine.

You can see exactly what this site knows about you and reset it at any time.