When sensitive data is already outside the boundary, deletion alone is not containment. I wrote about a response loop—discover, classify, neutralize, trace, learn—and where ClawBack fits: stateofsecurity.com/the-leak…

Could AI and robots help us grow old at home? I think practical home-care systems may be less than a decade away. My new essay explores the evidence and probabilities. notquiterandom.com/2026/10/0…

Prompt injection should change an AI agent's authority—not just create an alert. My new State of Security article defines four runtime states and shows how CaneCorso™ provides the content-risk control point. stateofsecurity.com/prompt-i…

I'm working on bringing my average day closer to my ideal day: movement, writing, learning, useful work, family, my dog, and rest. A life I don't need to escape. Here's the approach: notquiterandom.com/2026/10/0…

SSH honeypot logs still show the old password guesses. New account names, including claude and openclaw, are appearing too. Adam Hostetler examines what defenders can learn from the attempts. stateofsecurity.com/the-same…

In my recent passkeys article, I called out a point that deserves its own treatment: passkeys do not solve weak account recovery. Read the full article: Account Recovery Is Becoming the New Identity Attack Surface ▸ lttr.ai/AsUy9 #authentication #emergingthreats

Instead of saying, “X will happen,” it says, “Here's the range of what might happen, and how likely each scenario is.” That humility makes for much smarter risk management. Read more 👉 lttr.ai/AsTzU #investing #finance #BayesianMethods

The outcome is a much longer whitepaper, which you can download here if you are interested. Read more 👉 lttr.ai/AsTQn #crypto #blockchain #HoldingDigitalAssets

Who is in the materiality group?Typically: CISO, GC, CFO, CIO or CTO, privacy leader, communications, business owner, risk leader, and incident commander. Read more 👉 lttr.ai/AsRSw #Security #Infosec #Howto

Agentic Delegation: Moving beyond simple automation, attackers delegate entire phases of the attack lifecycle to semi-autonomous AI agents that plan, adapt, and persist across multiple days. Read more 👉 lttr.ai/AsQeW #Security #Infosec #AIAndMachineLearning

"The organization will transition workforce authentication from password-centric methods to phishing-resistant authentication using passkeys based on FIDO2/WebAuthn." lttr.ai/AsN9D #password #authentication #Security

Always know what AI agent has access to what... Read more 👉 lttr.ai/AsOwc #ai #aiagents #howto

One of the most important is a task and project management system I built called TaskGrid. Read more 👉 lttr.ai/AsOqL #mentalmodels #howto #productivity

"Account recovery, authenticator reset, passkey registration, passkey removal, device replacement, temporary access issuance, OAuth consent approval, and session revocation are security-sensitive identity lifecycle events." lttr.ai/AsOlm #authentication

It won’t magically make bad data good.It won’t remove all hallucinations.It won’t replace judgment. Read the full article: Building a Graph-First RAG Taught Me Where Trust Actually Lives With LLMs ▸ lttr.ai/AsMSU #llm #genai #ai

FIDO defines passkeys as FIDO authentication credentials based on FIDO standards, tied to an account, and used with the same process the user already uses to unlock a device. Read more 👉 lttr.ai/AsL8z #password #authentication #Security

Well, let’s just say we learned the hard way that pretty isn’t always practical. Read more 👉 lttr.ai/AsJHM #renovation #kitchen #bathroom

The model generally produces factually correct answers across multiple domains. Read more 👉 lttr.ai/AsHO6 #llm #lmstudio #genai

“If you cannot name who controls email, payments, and recovery, you do not have control.” #smallbusiness #cybersecurity #startup

Empirical studies also indicate that AI adoption is broad and growing... Read the full article: Your First AI‑Assisted Research Project: A Step‑by‑Step Guide ▸ lttr.ai/AsHHD #notebooklm #llm #research