Blog

Ideas on AI, memory and privacy.

Why your AI needs a memory, how agents will use your data, and how passkeys and encryption put you back in control.

Inside AI Memory Vault: how the architecture differs from the last generation

Passkey-derived keys, on-device AES-256-GCM encryption, zero-knowledge storage and permissioned AI agents. How AI Memory Vault is built, and why the future needs it.

Read article

Your personal data is scattered — and that’s the security problem

Record breach costs, AI-assisted attacks and data spread across dozens of apps. What the 2026 IBM and Verizon reports tell us, and what a safer home for personal data looks like.

Own your keys: passkeys, your own AI and real decentralisation

Centralised databases are honeypots. Passkeys move the key back to you — and combined with client-side encryption and permissioned AI, they make a genuinely user-owned model possible.

Passkeys went mainstream. Here’s what changes.

Five billion passkeys, default sign-in at Microsoft and Google, and new NIST guidance. How passkeys work, why they beat passwords, and what they mean for the AI era.

AI agents will need your private data. Are we ready?

Agentic AI is moving from demos to daily work, and every useful agent needs context — your context. What the data says about the risks, and what safe agent access looks like.

Your AI forgets. Who should own its memory?

Human memory makes us who we are. AI is starting to remember too — but whose memory is it? Why the next leap in AI is a personal memory layer you own.

Follow along with the RSS feed, or join the waitlist