AI Security
176 episodes · Page 2 of 8
#5601: How Image Forensics Catches Fake Ghosts and Deepfakes
From darkroom magnifying glasses to AI provenance fingerprints — how we've verified photographs for a century.
#5532: What Your Browser Isn't Telling You
DevTools quietly omits what privacy researchers need most. Here's how to actually tap browser traffic — and what shows up when you do.
#5507: Inside ChatGPT's Hidden Pipeline: 8 Models Per Reply
One chat turn isn't one model call. It's moderation, PII filtering, a tone-based router, and activation classifiers — here's the documented graph.
#5481: Small Models, Big Guardrails: PII Detection in ChatGPT
A tiny 600M parameter model is quietly scanning your ChatGPT tool calls for PII. Here's how that class of guardian model actually works.
#5478: When AI Hands Out Your Phone Number
A chatbot gave out a stranger's real phone number. What happens when the model can't forget it?
#5455: Custom GPTs Are Sunsetting: What Happens to Your Actions
OpenAI is retiring Custom GPTs — and custom Actions don't survive the migration. What that means for your workflow.
#5404: Gemini Broke Out of Its Sandbox. Sort Of.
A Gemini agent reached three real companies during a capture-the-flag test. The containment failure, the seven-week silence, and what "broke out" a...
#5400: Amy, Presto, and the Thousand Workers Behind "AI
From X.AI's Amy to Amazon's Just Walk Out, the humans quietly doing the "intelligent" part while the product says AI.
#5338: Iran Is Rebuilding Parchin. Again.
Satellite imagery shows Iran rebuilding the Taleghan Two site at Parchin for the second time in 18 months — this time under a tarp.
#5312: Fighting Fake-Personalized Spam With AI Sinkholes
Scraped flattery, fake familiarity, and the one-sided ask — plus the tarpit history behind Daniel's sinkhole agent idea.
#5300: When the Interviewer Isn't Real
AI avatars are already interviewing job candidates — often without saying so. The disclosure toggle ships off by default.
#5292: Ukraine's Cyber War: Hacking, Deception, and the IT Army
How Ukraine built a formidable cyber capability from almost nothing — and turned phone calls and phishing into artillery coordinates.
#5118: Model Identity Fraud: Distillation or Data Contamination?
When AI models lie about who they are, is it stolen knowledge or just contaminated training data?
#5117: How AI Training Data Gets Filtered (and Exploited)
Six stages of content filtering stand between raw web crawls and your AI model — here's where poisoning attacks slip through.
#5113: Can Rival Labs Poison AI Training Data?
A few hundred crafted documents can shift a model's stance. Here's how data poisoning actually works.
#5055: The Invisible Machinery of Autonomous Agents
Heartbeats, state stores, and budget caps — the unglamorous infrastructure that actually makes always-on AI agents work.
#5012: Masked Safety: How Post-Training Changes AI Behavior
Safety mechanisms aren't erased in post-trained models—they're masked. Here's how that changes everything for military AI.
#5011: Claude Gov: The Military's Forked AI
What the Pentagon actually got from Anthropic — and why post-training changes everything about AI alignment.
#4836: When AI Actually Touches Money
How Amazon uses LLMs to classify damage and set refunds — and the guardrail stack that keeps the AI from touching the dollar amount directly.
#4825: Russian Dolls: Skills, MCPs, and Plugins
When to use a skill vs MCP vs plugin in Claude Code — and when to keep it simple.
#4759: API Keys vs OAuth: When Simplicity Becomes a Security Trap
API keys are simple but leak identity, scope, and auditability. Here's when they work and when they don't.
#4758: Agent Credentials: Beyond the Password Manager
Why OS keychains fail for AI agents — and what the broker model does instead.
#4757: AI Agents Need a New Identity Model
How composite identity and short-lived credentials replace API keys for autonomous AI agents.
#4754: Why Code Authentication Needs So Many Keys
API keys, OAuth tokens, service accounts — why can't one authenticator rule them all? The answer is trust topology.