Read between
the lines.
AI security. System failures. Famous papers with uncomfortable questions. We read the research, inspect the assumptions, and argue about what actually holds up.
Enter the reading roomStart with a question.
Five papers, from classic system security to agentic AI. Each brief is deliberately short. The real work starts when you disagree.
05 briefs on the shelf
Not another
slide deck.
A simple way to turn reading into conversation. No expertise test. No need to have all the answers. Show up with one good question.
The setup
What problem was this paper trying to solve?
The breakdown
Understand the method, the evidence, and the claimed result.
The challenge
Pick apart assumptions, debate defenses, find failure cases.
The open question
What would we test next? Leave with a falsifiable idea.
CTRL+F is a space for reading security research actively, not collecting summaries. We pair foundational ideas with modern AI systems, asking what changed, what stayed broken, and what evidence would change our minds.
Papers are linked to their original sources. Briefs distinguish reported findings from our interpretation, and prompts are invitations to think, not tests with a single right answer.