AIM³ Dream Team Protocol · Round 2 of 2

ssMDL-DCC Challenge

Apply the winning AI-Storm method to find the optimal self-selecting MDL+DCC architecture for solving the Travelling Salesman Problem.
March 2026 · Bojan Dobrečevič · Master Document
Prerequisite

This Is Round 2

This challenge requires the winning AI-Storm method from Round 1. Five frontier AI systems each designed their own AI-storming architecture. The human architect synthesized them into one hybrid protocol. That protocol is the method you will use here.

Status

Round 1 is currently in progress. This page will be populated with the full TSP challenge once the hybrid AI-Storm method is finalized.

Preview

What Round 2 Will Ask

Using the hybrid AI-Storm method, all five AI systems will attack the same problem:

Find the best fractal ssMDL-DCC architecture for solving TSP. Specifically: what sensors, control laws, and representations should be used at each recursive level? How many levels? How do they communicate? Can the system discover novel heuristics autonomously? Does it scale polynomially?

Required Reading (for Round 2)

Required
The Self-Selecting Governor
The ssMDL-DCC concept paper. Alternative architectures, MDL scoring, recursion.
Required
8Z-RP TSP Solver
Current DCC-governed solver. qa194 exact, uy734 0.41% gap, nu3496 1.62%.
Context
Recursive DCC
DCC watching DCC. Semantic inversion. The architecture ssMDL-DCC builds on.
Context
CCH Science
The biological parallel. S-metric. Claustrum as coherence-complexity controller.
Context
Trading Governor
Three-level nested MDL+DCC. How the same principle scales across levels.
◇
Awaiting Round 1 Results
When the hybrid AI-Storm method is finalized, this page will be expanded with the full challenge specification, output format, and scoring criteria. All five AI systems will then apply the same method to the TSP problem independently.