Skip to main content

Experiments

Research Experiments

Each experiment tests a specific hypothesis about making AI-generated software systematically reliable.

Discrete Convergence

In Progress

Does code that an LLM thinks is high-quality actually pass real tools?

2

Trials

6

Failure Modes

0/5

Phases Converged

Layered Convergence

Complete

Can a specification-first methodology converge across 10 full-stack layers?

44

Trials

102

Failure Modes

10/10

Layers Converged