The 6 things AI
cannot replace.
Yscore is a new education benchmark. Not what a student knows — what a student is irreplaceably capable of. We grade the cognitive mechanisms that AGI can compute around but never embody.
6 questions · ~15 minutes · graded by Claude
The delta of mortality
AGI can compute, optimise, and predict with flawless logic. We don't grade any of that. Yscore evaluates only what AGI cannot do: biological intuition, physical reality, irrational serendipity, emotional resonance, and the existential weight of having skin in the game.
A 14-year-old with simple words and a 50-year-old CEO are on the exact same cognitive playing field. We don't grade vocabulary, grammar, or domain trivia. We grade the mechanism underneath.
The 6 Permanent Human Axioms
What we grade.
Intent
The Genesis of Purpose
AGI is a neutral engine — it lacks inherent desire or ambition. Intent is the strictly human capacity to feel friction in the world, cut through the noise, and architect the highest-level Why. The ability to say: "this is what matters."
Dynamics
Real-Time Biological Kinetics
AGI has no nervous system and cannot die. Dynamics measures the kinetic mastery of unscripted human friction in physical meat-space — where adrenaline, biological panic, and irrational motives dictate reality. Grace under fire, in real time.
Intuition
The Zero-Data Leap
AGI perfectly predicts the historical average. Intuition is irrational, un-digitized serendipity — the zero-data leap, based on lived human friction, that defies statistical logic but is proven right by human nature.
Discernment
Reality Anchoring
AGI hallucinates with mathematical plausibility, making articulate deception infinite. Discernment is the human auditor: spotting logical fallacies in highly articulate synthetic output and tethering theories back to gritty physical truth.
Taste
Combinatorial Resonance
AGI generates infinite perfection. But mathematical perfection is emotionally dead. Taste is the subjective, unquantifiable judgment required to curate vulnerability, shared culture, and deliberate imperfection.
Liability
Skin in the Game
AGI can calculate a 99% probability of success — but an AGI cannot go to jail, feel shame, or go bankrupt. Liability measures the uniquely human ability to bear existential weight and take accountability when consequences are irreversible.
How it works
6 scenarios. 1 Yscore.
- 01
Join through your school
Your school gives you a join code. Sign up, verify your email, and get access to the 6 scenarios.
- 02
Answer each scenario in your own words
No multiple choice. No rubrics about grammar. Write how you'd actually think or act. The scenario is the only thing that matters.
- 03
Claude grades you against the axiom rubric
Opus 4.6 with adaptive thinking reads your response and maps it to a 1-10 score against the specific axiom. You get two sentences of feedback explaining exactly what you did and why it landed where it did.
- 04
One fatal flaw can veto your Yscore
Any axiom scored at 3 or below doubles its weight in the weighted mean. You can't be brilliant at five axioms and be average at Liability — Yscore won't let you.
- 05
Your school sees the aggregate
Each school gets a rolled-up Yscore across all its students. The Y is the real education benchmark: how much of its student body the school has prepared for a world AGI can almost do everything in.
Your school is asking the
wrong question.
It's asking what you know. It should be asking what you can do that a machine never will.