Insights

What we think, and why.

Positions we hold because of what the work keeps showing us. Arguments rather than announcements. Each one is something we would defend in a room with your engineers.

Pieces: 4
Topics: AI, assurance, engineering
Cadence: When we have something
Author: The practice

01All pieces

Written by the people doing the work.

No ghostwritten thought leadership, no vendor talking points.

Your AI system is an architecture, not a model

Most AI security effort is aimed at the model. Almost every real failure we see happens somewhere else in the chain.

AI Security · 6 min

A finding is not a risk

Severity labels describe a vulnerability. They say almost nothing about what it would cost you.

Assurance · 5 min

Security that lives outside the flow gets routed around

If the secure path is slower than the insecure one, engineers will find the insecure one. Every time.

Engineering · 5 min

How to buy security testing without wasting the budget

The scoping conversation determines the value of the engagement more than the testing does.

Advisory · 6 min

Next step

Disagree with any of it?

That is a useful conversation.

These are positions, not doctrine. If your environment argues otherwise we would genuinely like to hear it.

Talk to a security expertSee our services