Security | Threat Detection | Cyberattacks | DevSecOps | Compliance

PSA: AI is NOT your safe space. That “quick question” you typed into an AI chatbot? It could be stored, reviewed, or used for training. Oversharing puts your data at risk. Protect yourself by keeping personal, financial, and work info out of chat prompts. Think before you prompt!

If you wouldn’t hand your house keys to a delivery driver, why hand your credentials to AI? In this Principles in Practice video, Anand Srinivas, VP of Product & AI at 1Password, explains a critical rule for secure AI use: Raw credentials should never be shared with large language models. Instead of sharing secrets, use them securely: Don’t send raw credentials over the data channel of a protocol like MCP Use proxies and secure autofill instead of sharing secrets Keep credentials out of prompts, embeddings, and fine-tuning data.

The Razorwire Christmas Party 2025 episode compares automation in law and cybersecurity, where junior roles shrink and the talent pipeline starts to break. AI pressure on tier one soc work in 2025 leaves new entrants with debt and fewer real training grounds, raising hard questions about the future of senior expertise.

Try THIS Prompt to Build a CRUD Note Taking App!

Can an under-the-radar AI tool actually build a secure, functional CRUD note-taking app from scratch? In this video, I put GitHub Raptor Mini to the test to see if it can design, implement, and reason through a real-world CRUD application — including authentication, data handling, and basic security considerations.