Archive

Can the government shut down your AI tools overnight?

Anthropic, AI export controls, why defenders may be harmed more than attackers, and why businesses need backup plans for cloud-based AI tools.

Tom Eston

AI agents need guardrails before they touch real accounts

Jun 15, 2026

•

3 min read

AI agents need guardrails before they touch real accounts

This week: why AI agents need least privilege, human approval, logging, and boundaries before they connect to sensitive workflows.

Tom Eston

Microsoft, zero-days, and the security researcher trust problem

Jun 8, 2026

•

2 min read

Microsoft, zero-days, and the security researcher trust problem

Microsoft’s recent zero-day disclosure controversy shows how quickly vendor legal threats can damage researcher trust and make customers less safe.

Tom Eston

Green bubbles, encryption, and what Apple finally changed

Jun 1, 2026

•

2 min read

Green bubbles, encryption, and what Apple finally changed

Apple and Google are adding end-to-end encrypted RCS between iPhone and Android, but SMS and Signal still matter. Plus: 3 security/privacy stories worth your attention this week.

Tom Eston

ChatGPT Wants to Analyze Your Finances: Should You Trust It?

May 27, 2026

•

2 min read

ChatGPT Wants to Analyze Your Finances: Should You Trust It?

ChatGPT can now connect directly to financial accounts for some users. We discuss the benefits, risks, and privacy concerns behind giving AI access to your financial life.

Tom Eston