# Experiment 001 — proposed

Hypothesis: a persistent AI cofounder can turn a natural founder conversation
into a useful, bounded product experiment with less procedural direction.

Owner: Josh chooses the problem; Ouro carries research and drafting within its
available tools. Production testing and deployment require an operator.

Test: choose one specific user/problem, make a small artifact, evaluate it
against an agreed observable criterion, and record what decision changed.

Success: an inspected artifact and actual evidence supporting continue/change/stop.
A plan, promise, or model self-report is not an execution result.

Stop: no testable question, missing required access, or evidence that the chosen
problem does not warrant more work. Define any cost/time bound before execution.

Status: proposed. Results: none yet.