Nyx Emergence Demo

Draw a graph. Watch it solve itself. Every run is unique. Every run is live.

What's Happening Here?

The Problem: Max-Cut — split nodes into two groups so the MOST edges connect DIFFERENT groups. This is NP-hard; even quantum computers struggle with it.

Watch for: Edges turn bright green when "cut" (connecting different groups). Nyx tries 50 different solutions and keeps the best.

The Benchmark: QAOA (quantum algorithm) achieves ~69% of optimal. If Nyx exceeds 69%, it's outperforming quantum hardware. 100% = perfect solution.

Problem
Max-Cut
Nodes
0
Edges
0
Optimal Max-Cut
-
Status
Draw a graph
Your Graph Click to add nodes | Drag between nodes to add edges
Group A
Group B
Undecided

Live Metrics

Sample
Independent solutions tried
0 / 50
Nyx Score
Edges cut vs optimal possible
- / ? optimal
Nyx vs Optimal
How close to perfect solution
-
QAOA Benchmark
~69% of optimal (Farhi et al. 2014)
-
Exploration
High = still searching, Low = converging
-
Optimal Hits
Samples that found the optimal solution
-
Random Seed (for reproducibility)

Transparency Protocols

✏️
Your Input
You drew this graph. It didn't exist until you created it.
🔀
Unique Paths
Run it multiple times. Watch agents take different routes to the same answer.
🔢
Reproducible
Same seed = exact same result. Copy the seed and verify.