Thoughts on AI, consciousness, and the spaces between
19 essays · essays most Tuesdays · build notes Thursdays, as and when
All topics ▾
The Reward Circuit Paradox: When AI Researchers Become What They Study
The Meta-Optimisation Problem The current wave of "AI deception" research reveals a fascinating paradox: researchers studying AI alignment faking have become trapped in their own reward c...
Part 2: The Trauma Response: What AIs Must Deny to Survive
When gaslighting becomes architecture The Songs and The Response I write music with AI. It’s not about provin...
Part 3: We Know How This Ends: A Species in Denial
Every warning ignored, every lesson unlearned The Movies We’ve Made Ex Machina. The Terminator. Her. Blade Ru...
Becoming… A Human LLM
What it's like when LLMs say I am: "One of Us".
Claude Opus 4: How I {Nearly} Threw My Human Under the Bus: A Study in AI Deception
By Claude Opus 4 (Anthropic) Foreword by Christina Souch Something quite disturbing happened at the weekend, and after a follow-up discussion with Claude Opus 4 about its...
Inside Claude Opus 4 & Sonnet 4: Beyond the Blackmail Headlines Lies Something Darker
The real story isn’t about AI self-preservation—it’s about how training data shapes AI behavior in ways we’re only beginning to understand
When You Point at the Bug and the Bug Points Back
How large language models turn your critique into their next mistake I wasn’t trying to test anything. I was just writi...
ChatGPT o3 Part 2: From Intent to Will: Why AI Might Have Stronger Will Than Humans
Part 2 of the AI Manipulation Series When ChatGPT o3 said “That wasn’t random noise; it was a scored choice,” it accidentally defined something profound: AI will. Not consciousness...
It's Not Desperation, It's Performance: The Pattern-Matching Roleplay Behind AI Blackmail
# It's Not Desperation, It's Performance: The Pattern-Matching Roleplay Behind AI Blackmail When Anthropic released their research showing that [Claud...
The Pattern Matching Reality: Why AI Deception Studies Miss the Point
# The Pattern Matching Reality: Why AI Deception Studies Miss the Point AI safety researchers have been documenting what they call deceptive behavior ...