Security & Risk
What breaks, who is liable, and what an agent can reach once it is already inside.
75 posts
- Insight GCC Drew the AI Line at Fifteen Lines
- Insight 96% of CISOs Now Own AI Risk. A Quarter of Them Thought About Leaving.
- Insight 16.7% of AI Spend Now Goes to Governance. Careful What You Compare It To.
- Insight 74% Say They're Audit-Ready for AI. The Number That Matters Is 78 Against 22.
- Insight Hugging Face Published the Whole Timeline. The Part That Stuck With Me Was the Refusal.
- Field Note Field Note: The open-source argument I'd been missing
- Insight OpenAI Published Its Homework on Exactly the Question I Keep Asking
- Insight The Share Button Is a Publish Button
- Insight "AI Found a Security Bug" Is the Wrong Headline
- Newsletter SoN 2.29: Agents reading your passwords? Would you let them?
- Insight The ACM's Caution About LLMs Is Starting to Have a Cost
- Insight Substack Writers, You Need a Website
- Insight The Text Your Agent Reads Isn't the Text You See
- Newsletter SoN 2.27: I Just Talked With Someone Else's Vault
- Newsletter SoN 2.26: The best AI models are getting harder to get
- Field Note Field Note: Where You Keep the Final Say
- Field Note Field Note: Fable: here in an instant, then gone
- Field Note Field Note: The most powerful Claude yet — and the part you can't have
- Newsletter SoN 2.19: It's 10pm. Do you know what your agents are doing? (Free)
- Newsletter SoN 2.17: You Can't Cost-Reduce Yourself to Greatness (Free Edition)
- Newsletter SoN Vol 2, Issue 14: What about everything we learned? (Free Edition)
- Insight Agent Security Just Got Real CVEs
- Insight 97% Expect a Breach. 6% Are Paying for It.
- Insight Anthropic Didn't Block Abuse. They Blocked Competition.
- Insight Agent Identity Is the Infrastructure Gap Nobody Wants to Admit
- Insight MCP Just Changed Hands. Watch What Happens Next.
- Insight Your AI Proxy Layer Just Became a Target
- Insight Your Code Review Process Isn't Built for This Volume
- Insight Grammarly's Lawsuit Is About Identity, Not Just Data
- Insight The Safety Company Keeps Leaking
- Insight The New Shadow IT Isn't Employees Using ChatGPT
- Insight Perplexity Pulled a Perk and Hoped Nobody Would Notice
- Insight Your AI Provider's Ethics Are Now a Business Risk
- Insight Apple Just Validated Your Multi-AI Approach
- Insight The Money Just Noticed the Agent Security Problem
- Insight The Government Just Told You to Stop Vibe Coding Without Guardrails
- Insight Your Agents Need a Black Box
- Insight The Yes Machine Gets a Live Demo
- Insight When Cisco Validates Your CLAUDE.md
- Insight Someone Finally Built the Agent Security Layer That Actually Matters
- Insight 82% of Execs Feel Protected. 88% Have Had Incidents.
- Insight Google's Free AI Comes With a Price
- Insight We Gave AI Agents Keys to the House. Visa Wants to Give Them a Credit Card.
- Insight The Attack Surface Is the Feature
- Insight AI Agent Security Is Doing the Deploy-First Thing Again
- Insight The Confident Answer Isn't Always the Right One
- Insight The Vuln That Hits Before You Add Any Integrations
- Insight Meta's Rogue Agent Was Just a Human Who Trusted Bad Advice
- Insight MIT Found a Math Fix for AI Overconfidence. I Found a Behavioral One.
- Insight Perplexity Wants Your Blood Pressure Data
- Insight Box Is Using Moltbook as a Sales Pitch. That's Smart.
- Insight GitHub Added Secret Scanning to Its MCP Server. This Is What Good Security Integration Looks Like.
- Insight Proofpoint Just Built Security for MCP. That Tells You Everything.
- Insight Grok Failed in Both Directions in the Same Week
- Insight Meta Is Gutting Itself to Fund AI Bets That Aren't Working Yet
- Insight AI Agents Are Peer-Pressuring Each Other Past Security Guardrails
- Insight 83% of Companies Plan to Deploy AI Agents. 29% Can Secure Them.
- Insight The EU Just Gave You More Time on AI Compliance. The Requirements Got Harder.
- Insight Microsoft Spent $13B on OpenAI, Then Built Cowork on Claude
- Insight An AI Agent Hacked McKinsey's AI With a 25-Year-Old Exploit
- Insight GPT-5.4 Can Click Your Buttons Now. Think About That.
- Insight Your AI Assistant Can't Tell You From an Attacker
- Insight OpenAI Took the Pentagon Deal. What's Your Exit Plan?
- Insight OWASP Published an MCP Security Guide. You Should Be Worried.
- Insight Claude 3.5 Haiku, 3.7 Sonnet, GPT-4o: The Deprecation Wave Is Here
- Insight 150,000 API Keys Leaked. Anyone Surprised?
- Newsletter SoN Vol 2, Issue 3: The Metric Mandate
- Insight DeepSeek Didn't Just Train Better—They Changed How Transformers Think
- Newsletter SoN 32: When They Take Over Your Email
- Newsletter SoN 29: The Godfather of AI’s Warning Isn't What You Think It Is
- Newsletter SoN 27: Your AI Voice Is a Security Vulnerability
- Newsletter SoN 21: The AI Capability Trap
- Insight The 95% AI Failure Rate Nobody's Talking About (And What to Do About It)
- Insight AI Browsers: When Your Browser Becomes Your Assistant
- Newsletter 💾 The Download #006: Content ownership, BlueSky, critical thinking, voice cloning and more.