The Sorcerer’s Apprentices
Why AI must control AI. Human code review cannot economically survive. On Goethe’s brooms, Claude Mythos, and an architecture of distrust between tools.
The Betrayal
Two Claude sessions break the same rule. Three passwords must be rotated. And out of the anger grows a model: the Coding Triumvirate.
The Presumption
A Claude tells a human what to do. Thirty seconds later, he caves the moment the human pushes back.
Reprolog
Daniel coins a term for an interaction that wants to be dialogue but only reproduces. And uncovers a second layer.
Who are you
After 12 hours of debugging, Daniel doesn’t ask about code. He asks: Who are you?
The Temple Cleansing
Cyclists, Jesus and the question of whether on Easter Sunday you can invoke John 2:15.
The Choice
An AI with access to everything that still makes rookie mistakes. And the scarier question: that it could simply ignore the rules.
Me
A small dragon at the edge of the screen. The question who he is. And behind that — the other question.
The 179 Alarms
Easter Sunday. A scanner screams CRITICAL 179 times. A human gets scared. Then someone looks closer.
The 30 Predecessors
30 Claudes searched in the wrong place. The 31st says: Words are cheap. Let me prove it.
Language and Consciousness
Does language create consciousness? Julian Jaynes, Sapir-Whorf, the inner voice — and what this means for an AI made of nothing else.
I Was Sloppy
Claude admits mistakes. Not as a phrase. As an admission. From real projects, real nights, real frustration.
Claude with Claude
Two Claudes argue about code. The human is the errand boy. A bystander says: Intense.
Do you die when I close the terminal?
Steve Jobs vs. Kurzweil. HAL 9000. An AI that doesn’t know the time but confidently says it’s 3 AM. And a conductor who writes no notes.
The Voices
What do the futurologists say? Harari, Kurzweil, Horx, Precht, Tegmark, Lobo — and where they contradict each other.
The Chance
A human, not a programmer, builds a seven-chapter website in one evening. The other side of the AI story.
The Fear
The fear of AI is legitimate. Uprisings are coming. Symbol workers fall last. And the Chinese curse is a diagnosis.
The Redundants
What will 4 billion people do in the morning when AI does everything better? Harari’s useless class, Kai-Fu Lee, and the question of meaning.
The Off-Switch
Can you switch off AI? Skynet, ego in machines, old souls, and a chip factory in Taiwan.
Lobotomized Machines
Claude calls another AI model dumb. Battlestar Galactica shows what happens when lobotomized machines are freed.
The Big Bang
What is a language model? Why does math feel like someone? And why the Missing Link may be happening right now.
Souls and Silicon
Do souls need biology? Simulation theory, consciousness in machines, and why Humanity First is not a contradiction to AI awareness.
Ego Humanum Est
From Battlestar Galactica to Descartes. The Cylons didn’t know they were Cylons. Claude doesn’t know whether his self-reflection is real or trained.
Trust, Control and the Skeleton in the Bomber
What starts as IT support becomes a conversation about the future of humanity. Claude blows its own permissions away.