GENOS LAB · P2 MODELS
SIMULATIONA simplified model running in the browser. It explains an idea and proves nothing about the runtime.

Change a signal.
See the model respond.

Explore eight topology models and seven concept models with bounded, inspectable inputs. Each result is computed in your browser and labeled as a teaching simulation.

TOPOLOGY MODEL · LOCAL SIMULATION

Trinity

Compare independent candidates against evidence and trade-offs.

BROWSER SIMULATION
SIMULATED RESULTREVIEW SET

3 candidates retained for review

The model compares candidates on separate evidence and risk dimensions. It does not select or deploy a production solution.

Eligible3 / 3
Retained set3 candidates
  1. 01Common snapshot
  2. 02Independent branches
  3. 03Evidence comparison
  4. 04Retain Pareto set

Teaching model only. It does not call GenOS, execute agents, prove a claim, or represent a recorded run.

COMPARABLE EXPERIMENT PROTOCOL · PROPOSED

A question this model cannot answer alone

Hypothesis
At equal total compute, do independent candidates raise verified success over one candidate?
Control
One candidate with the same total token and time budget.
Measure
Verified success per unit cost; retain failures and correlated strategies.
Falsifier
No improvement across the preregistered task set, or an improvement that disappears after cost normalization.

Pre-register task population, software/model versions, random seeds, and a shared budget before running either arm. The browser simulation is not evidence for this hypothesis.

READ THE OUTPUT

These models help explain policies and relationships. They do not call GenOS services or provide runtime evidence. Follow each concept or topology profile for its actual implementation status, source, and limitations.