Blog
Welcome to our blog. Here, Edgeless Systems employees write about all things Confidential AI.
When 90% means 90%: calibrating one-token LLM decisions
Raw one-token probabilities are overconfident. One temperature fixes most of it without labels. About 100 labels also make more answers correct and put an error bound on the answers you automate.
OpenAI Private Intelligence: what it promises, and what you can verify today
OpenAI now says AI inference belongs in confidential computing. What the DevDay announcement contains, what it leaves open, and three questions that separate a verifiable property from a promise.
Turn GLM-5.3-Flash into a Jev-like System One model
Typed decisions in a single forward pass, about as accurate as Jev and faster from Europe. With a playground and a benchmark on 29 datasets.
Confidential containers in Kubernetes: what a real operator exclusion requires
A technical operator exclusion in Kubernetes is possible today, but only one architecture actually delivers it. The short version, plus where to read the full paper from the 21st German IT Security Congress.
Building AI services for professionals bound by confidentiality obligations: a guide with Privatemode AI
How to use LLMs productively without the provider accessing the data – confidential computing and the confidentiality agreement under § 203 StGB that technology can't replace.
Apple Private Cloud Compute moves to Google Cloud
Apple Private Cloud Compute now runs on third-party hardware. As confidential AI becomes standard for private inference on widespread data-center hardware, the real question is about transparent verification.
AI Risks: BaFin and DORA on the secure use of LLMs
Why Confidential Computing is a technical response to the core ICT risk of cloud-based AI.
Why a third-party confidential-computing layer is required
Even when your cloud provider offers confidential computing, gaps remain. Here's what they deliver, where they stop, and why Contrast and Privatemode AI complete the picture.
Privacy-preserving LLM inference
Learn how privacy-preserving AI inference protects sensitive data when using LLMs, from cryptographic approaches to redaction.
C5:2026: Germany’s BSI makes confidential computing and remote attestation a standard
Germany’s BSI C5:2026 catalogue introduces mandatory criteria for confidential computing and remote attestation. See how Privatemode AI already meets the highest attestation standard.
Confidential AI in clinical care
How AI-powered documentation is reducing the administrative burden in healthcare and why Confidential AI is the answer for privacy and scalability.
Confidential RAG: secure AI knowledge without data leaks
Deploy AI knowledge management for confidential data without exposing information to third parties. Enterprise guide to secure RAG, local inference, and privacy.
AI memory features get baked into browsers – We build privacy back into them
AI browsers are racing to introduce 'memory' features, but they raise major privacy risks as browsing context gets sent to the cloud. Privatemode offers a safer path using Confidential Computing.
Secure prompt caching for fast AI inference
How to build fast AI inference with prompt caching while keeping your prompts secure.
Is ChatGPT safe?
Many businesses hesitate to use ChatGPT for sensitive data due to privacy concerns. While OpenAI claims to implement strong security measures, its privacy policy lacks transparency, especially on data sharing.