AI & Human Creativity Horizon Simulator

5-Stage Agency Matrix Human vs AI Share

PRESET PARADIGMS

1. Seed Ideation Human Lead

Human Ratio 80%

Exploration Temp 0.9

2. Draft Expansion AI Velocity

Human Ratio 20%

Exploration Temp 0.6

3. Counter-Intuitive Search Breakthrough Leap

Human Ratio 50%

Exploration Temp 1.2

4. Aesthetic Critique Taste Curation

Human Ratio 90%

Exploration Temp 0.2

5. Final Polish & Voice Artisanal Touch

Human Ratio 70%

Exploration Temp 0.4

2D Semantic Novelty & Cliché Basin Map Preset: Adversarial Co-Creation

Cliché Basin

Human Synthesized

Move 37 Breakthrough

Avg Distance to Cliché Basin

4.82 σ

Variance / Spread

32.4

Deskilling Risk

Low (24%)

Horizon Telemetry N = 100 Iterations

Novelty Score
84.5

Homogenization Index

18.2

Human Agency Index

62.0

Breakthrough Rate

14.0%

Assessment Synthesis

High-Leverage Hybrid: Avoids statistical mediocrity while expanding exploratory search boundaries beyond conventional genre tropes.

💡 Grounded Quora Source Insights

AlphaGo Move 37 Principle: AI when directed into non-standard search spaces produces novel, counter-intuitive moves humans never imagined.
Probability Cliché Trap: LLMs default to median training data, causing artistic homogenization without active human aesthetic critique.
Deskilling Friction: Outsourcing seed ideation and critique erodes human artistic taste curation over time.

Creativity simulation: weights, preset overrides and generated points

Your browser does not support video.

Read the explanation

The pure-automation preset averages human allocation across five stages. Its seed ratio is point zero five and the other four are zero, so average human agency is one percent. Homogenization uses one minus forty percent of seed allocation plus sixty percent of critique allocation, times ninety. With critique zero, this gives eighty-eight point two. Bars use six pixels per index point. These are authored scalar weights. They do not inspect creative output, observe an artist, run an LLM, measure training-data concentration or establish deskilling. The program labels a synthetic workflow score; its causal arithmetic is transparent but its interpretation requires independent evidence. Under the general formula, co-creation search temperature one point two contributes forty-two, seed allocation point eight contributes twenty-four, and homogenization twelve point six contributes twenty-six point two two. Rounded novelty would be ninety-two point two. But the exact adversarial co-creation preset returns fixed novelty eighty-four point five and homogenization eighteen point two before executing that formula. Bars compare general-formula novelty and the override at six pixels per point. Moving a slider clears the preset identity and returns to dynamic arithmetic. Therefore a numerical jump can reflect the fixture exception rather than a measured improvement in originality. The video does not present the labels as an empirical creativity score. The general breakthrough rule multiplies search temperature by eight and adds counter-intuitive human allocation times six, then rounds and clamps. Pure automation yields four percent. The co-creation fixture instead returns fourteen percent. The scatter plot always generates one hundred synthetic points and marks the first rounded percentage count as breakthroughs before choosing their coordinates. Bars compare four and fourteen assigned points at forty pixels each. No creative samples are evaluated and no novelty is discovered by the plot. Re-simulate changes synthetic coordinates, while the exported telemetry remains rubric arithmetic. All original source functions and current footer content must remain intact when the explanatory video is placed below the tool.