A builder can prove that a door opens.
The hinges are installed. The frame is square. The handle turns. The door passes inspection.
These are meaningful facts.
If the door leads to a staircase no wheelchair user can reach, another question remains. If the building was meant to become a public meeting place and nobody can enter, the work has not met its human purpose.
In the second week of September, a recurring design question came into focus for us: what does proof actually prove?
Several truths can sit beside one another
A computer test can prove that a function behaves according to a specified rule. A delivery receipt can prove that a message reached a service. A signed approval can show that someone authorized a particular action.
None of those facts automatically proves that a person's original goal was achieved.
The distinction is important precisely because these forms of evidence are valuable. Each answers a different question.
Did the tool work?
Was the action attempted?
Was the action permitted?
Did the intended change happen in the world?
A well-run system needs to keep these answers distinct long enough for people to make good decisions.
The danger of a green light
Many workplaces have learned to worship the completed checkbox.
A report is sent, so the assignment is marked finished. A site is launched, so the project closes. A meeting is held, so the stakeholder has been consulted.
The status changes color. The underlying need may remain.
Machines will make it easier to generate passing checks and polished deliverables. This increases the importance of choosing the right thing to measure.
Verification can establish that a promise was followed. An outcome asks whether the promise was worth making and whether it came true.
Consider a school that installs a new system for booking parent meetings. It may work perfectly in a test. The real measure includes whether families can understand it, whether parents with limited access can participate, and whether the conversations actually take place.
Those are matters for observation, people, and honest feedback.
Ask for the next layer of evidence
The next time someone announces that a project is complete, ask which claim the evidence supports.
If the evidence shows delivery, celebrate delivery. If it shows participation, say participation. If it shows an improved result for a person, say what changed and how you know.
Where the answer is uncertain, keep the uncertainty visible.
That practice does not slow ambition. It gives ambition a chance to meet reality.
The green light should belong to the right question.