Richard

This reshapes how you think about privacy, what you share online, where you're actually protected, and what defenses still work.

You still have the right to anonymity. You just can't exercise it anywhere that matters.

Age verification doesn't ban anonymity. It makes it irrelevant.

How many of your controls could you remove tomorrow without changing your actual risk posture?

The organizations I worry about aren't the ones with fewer controls. They're the ones where control count keeps climbing but nobody can explain what problem each one actually solves.

The more controls I see, the more worried I get. Dense control environments don't always mean secure organizations. Sometimes they mean nobody trusts the people enough to let them make decisions.

Here's the article I wrote on it. Why AI security testing looks nothing like traditional pentesting, and what you should actually be measuring: trustboundary.ghost.io/why-ai-securit…

You approved the AI deployment. Did you test it right? Prompt injection appears in 73% of production AI deployments. It's the number one vulnerability in OWASP's 2025 LLM Top 10. And most pentest teams miss it because they're testing against the wrong threat model.