Tutorial 07 · Beginner
Validate a Problem Worth Solving
Turn one founder problem claim into a traceable evidence brief and choose the smallest useful test without mistaking public interest for demand.
You will build
A problem evidence brief, assumptions register, dated source log, run receipt, and one bounded next validation test
You will strengthen
Separating evidence from enthusiasm, challenging weak signals, and deciding what must be tested next
Learning map
Tools and capabilities
Tools
Capabilities
Before you begin
What you need
- Codex in the ChatGPT desktop app or another Codex surface
- A local folder where you can safely create the Founder Operating Workspace
- One real venture problem claim, or the provided fictional Sample Venture research
Practice assets
Download what you need
Step-one preview
Frame the problem decision
Objective
Define one customer problem claim, the founder decision it must inform, and the evidence that would support or weaken it.
Try it now · In your Founder Operating Workspace and Codex Plan mode, or a normal Codex conversation
- Download and extract the Founder Operating Workspace into a dedicated local folder, then read START-HERE.md, AGENTS.md, ROUTING.md, and the validation METHOD.md before asking Codex to change anything.
- Choose your real venture when you can share a bounded, non-sensitive problem claim; otherwise choose The Sample Venture now and use its fictional research instead of inventing evidence.
- Write the claim as: When [specific segment] is in [situation], [problem] creates [observable consequence].
- State one decision this run may inform, such as whether to conduct five interviews, and state that it does not authorize building, purchasing, contacting, publishing, or sending.
- If Plan mode is available, open it with /plan or Shift+Tab and ask Codex to challenge the claim and plan this evidence review; otherwise ask for the same numbered plan in a normal conversation and do not approve file changes yet.
Visible result
You can see one bounded problem claim, one decision boundary, criteria that could support or weaken the claim, and a plan that has not changed any files.
Completion check
I chose the real or fictional route, framed one falsifiable problem claim, and kept consequential actions outside this run.
If you get stuck
Narrow the statement to one segment, situation, problem, and consequence; if a required fact is unknown or sensitive, mark it unknown and switch the entire run to The Sample Venture before continuing.
I’m stuck on this step