Security Outgunned: Measuring Software Defined Attack Surface
summary
A quick method applies basic predictive analytics (Approximate Bayesian Computation) to POC results from two vulnerability scanners: one reports 30 critical findings, the other 37, then runs R code and 100,000 simulations to merge priors with observed data. The model estimates error rates near 33% vs 34%, produces credible intervals, and multiplies simulated false-positive/false-negative hours and fully loaded costs to forecast annual impacts (examples shown ≈ $148k vs $164k and larger-scale $372k vs $459k), helping pick the lower-impact tool.
AI-written summary · about
chapters (8)
- 0:00Introduction
- 3:40Believe Therefore I Measure
- 8:42Believe Therefore I Mush
- 14:45Data munging
- 15:00Model
- 17:30Visual Comparison
- 17:41Results
- 21:12Summary
related talks
39mThe AI Arms Race in Vulnerability Management, Who’s Winning?
21mState of Attack Surface Elimination in a World Plagued by Vulnerable Software
54mTowards Real World Cyber Risk (Panel)
48mMonitoring Application Attack Surface and Integrating Security into DevOps
53mAI-equipped Threat Actors Versus AI-enhanced Cyber Tools: Who Wins?
44mOpening Keynote: Prove It! Confronting Security With Data.
43mThe Art of Vulnerability Management - Practical Application of Cyber Triage
50mMeasuring Adversary Costs to Exploit Commercial Software...