The Session You Cannot Take With You | EARENDIL

Lumo: Privacy-first AI assistant where chats stay confidential

Meet Lumo, the zero-access encrypted AI assistant by Proton that does not track or record your conversations. Ask me anything — it's confidential

Lumo: Privacy-first AI assistant where chats stay confidential

Meet Lumo, the zero-access encrypted AI assistant by Proton that does not track or record your conversations. Ask me anything — it's confidential

Making end-to-end encrypted AI chat feel like logging in

We want private AI chat to be simple. Yet today, many end-to-end encrypted experiences still have a level of friction that make them feel like they’re from another era: it usually either involves a long seed phrase users are asked to “store securely,” insecure password based encryption, or apps that aren’t cross-device and lose your data periodically (on reinstall, browser cache clear, etc).

Confessions to a data lake

I’ve been building Confer: end-to-end encryption for AI chats. With Confer, your conversations are encrypted so that nobody else can see them. Confer can’t read them, train on them, or hand them over – because only you have access to them.

Prompt Caching In Agents | EARENDIL

How prompt caching shapes the cost, latency, tools, and architecture of coding agents, and what Pi does to keep cache behavior visible.

Unlinkable Inference as a User Privacy Architecture

Unlinkable inference is a technique that provably sandboxes your AI requests from each other and from your identity. We discuss its building blocks, applications, and how it fits into the broader landscape of private personal intelligence.

OpenAI unveils Lockdown Mode to protect sensitive data from prompt injection attacks | TechCrunch

Even with Lockdown Mode, ChatGPT could be vulnerable to prompt injections, but the goal is to reduce the likelihood that sensitive data gets shared in the process.

Values | EARENDIL

Earendil is a public benefit corporation crafting software and open protocols to strengthen human agency, bridge division, and cultivate lasting joy.

Shlok Khemani on Twitter / X

Super interesting that eve doesn't ship with any long-term or cross-session memory (yet). The main context-management lever out of the box is conversation compaction. This is unlike OpenClaw and Hermes, each of which comes with a default opinionated memory implementation. https://t.co/DDIyc4rFtW— Shlok Khemani (@shloked) June 17, 2026

Gajesh on Twitter / X

TL;DRapple has turn on this switch for everyone to participate in decentralized inferenceppl can rent out their unused compute space and anyone can use this with privacy guarantees https://t.co/LTP4zyjsdt pic.twitter.com/8Dvo7XK8jJ— Gajesh (@gajesh) February 18, 2026

Confidential Inference via Trusted Virtual Machines

Announcing a new collaborative research paper on Confidential Inference, a set of tools to improve the security of our model weights and of our users' data

Architecture

Overview of Ente's end-to-end encrypted architecture—learn how your data is encrypted on-device, securely shared, and safely stored using cryptography.

Top 4 AI chatbot privacy concerns and how to mitigate them | TechTa...

Explore four key chatbot privacy concerns, as well as privacy protection strategies for individual users and organizations hoping to safeguard user data.

Document: requestStorageAccess() method - Web APIs | MDN

The requestStorageAccess() method of the Document interface allows content loaded in a third-party context (i.e., embedded in an <iframe>) to request access to third-party cookies and unpartitioned state. This is relevant to user agents that, by default, block access to third-party, unpartitioned cookies to improve privacy (e.g., to prevent tracking), and is part of the Storage Access API.