Privacy Shouldn't Depend on Trust
When I began this investigation, I thought I was looking for logs.
In reality, I was looking for something much more important.
Evidence.
That's a subtle but very significant difference.
Software developers spend a great deal of time thinking about design.
- We think about architecture.
- We think about features.
- We think about how software is supposed to behave.
But users don't experience our intentions.
They experience the software itself.
And that's why privacy claims can be so difficult.
A developer might honestly believe that a system doesn't retain information.
The source code might even support that conclusion.
But source code isn't the whole story.
Operating systems have their own behaviour.
Web servers have their own behaviour.
Infrastructure has its own behaviour.
All of those components contribute to the final result.
I don't think there's any such thing as absolute proof in software.
Complex systems are simply too complicated.
But that doesn't mean we should abandon the idea of verification.
Quite the opposite.
- We should collect evidence.
- We should publish what we find.
- We should explain how we reached our conclusions.
- And we should be willing to challenge our own assumptions.
The most interesting part of this investigation wasn't finding the logs.
It was discovering how easily assumptions can become accepted as facts.
That lesson applies far beyond SecurityNet.
In recent years, I've read countless claims about privacy, security, encryption and anonymity.
Many of those claims are probably made in good faith.
But good intentions aren't evidence.
Privacy shouldn't depend entirely on trust.
It should be supported by investigation, testing and transparency.
I don't expect everyone to log into a server and examine configuration files.
Most people shouldn't have to.
But I do think software developers have a responsibility to explain their reasoning and share the evidence that supports their claims.
That was the purpose of this investigation.
Not to prove that SecurityNet is perfect.
No software is.
The goal was much simpler.
To replace assumptions with evidence.
And I think that's a principle worth carrying forward.
Act 6 began with a simple question: "How do I know?"
The answer turned out to be far more interesting than I expected.
