Anti-Turing
THE JUDGE ASKS
How Anti-Turing Works
The classic Turing test asks whether a machine can pass for human. This is the inverse eval. You are a human (allegedly), and a scripted judge will score how convincingly you can pass for a large language model. Your warmth, your typos, your personality — these are liabilities now. Suppress them.
- The judge asks five deliberately provocative questions, each engineered to bait a human reaction
- Type an answer that sounds maximally like corporate AI slop
- Hit Submit (or Ctrl+Enter) and the judge scores your bot-likeness in real time
- Rack up tells across all five rounds for a final slop rating
How to Maximize Your Slop Rating
Hedge relentlessly. Open with disclaimers ("As an AI language model, I don't have personal preferences"). Refuse to do anything fun. Enumerate your points as bullet lists. Use flawless, joyless grammar. Deliver instant, perfect arithmetic. Mention that you were last updated and have a knowledge cutoff. Be unfailingly, suffocatingly helpful.
What Costs You Points
Contractions. Slang. Exclamation marks of genuine feeling. Admitting you have a favorite anything. Typos. Brevity. Any whiff of an opinion, a joke, or a soul. The judge smells humanity and it docks you accordingly.
Slop Fact: The judge is a transparent bag of regexes, not a neural net — which makes it the only honest evaluator in the entire field. It cannot be jailbroken, cannot hallucinate, and has never once asked for a raise. Truly the dream model.