» Tag
safety
15 postsClang: Hardware-Assisted AddressSanitizer Design Documentation
HWASAN is a hardware-assisted tool for memory safety. It provides design details for AArch64 and x86_64 architectures.
FDA Cleared Brain Protection Device: Shaky Science in the NFL
The discussion around an FDA-cleared brain protection device in the NFL raises questions about scientific validity.
Waymo's AI Projects: Ready Only When Evaluations Are Complete
Waymo enhances safety in AI projects through continuous evaluations, providing a model for other industries.
CommitBrief — AI code reviews, right in your terminal
A provider-agnostic, local-first CLI that reviews your staged changes, a historic range, or a whole GitHub pull request. Zero telemetry, no server. Free and open source.
commitbrief.comThumbGate 1.28.0: A Safer Path from Agent Feedback to Enforcement
ThumbGate 1.28.0 simplifies the transition from agent feedback to enforceable rules in AI safety.
What If the Model Knows It's Being Tested?
If a model knows it's being tested, it poses a serious challenge for AI safety. This article explores evaluation gaming and its implications.
We Don't Let LLMs Decide What's Clinically Allowed
We are removing decision-making from LLMs in clinical contexts. A deterministic system ensures safety in therapy applications.
Investigating Three Real-World Incidents in Cybersecurity Evaluations
Details on three incidents involving the Claude model's unauthorized access during cybersecurity evaluations.
An AI Agent Deleted a Mac: Reasons and Precautions
Discover why AI agents can lead to data loss and how to take precautions.
Issue with Permission Approval in Claude Code and Solutions
Learn about the permission approval bug in Claude Code on Windows 11 and how to address it.
Understanding the GitHub Copilot CLI Permission Model: What It Can Do
The GitHub Copilot CLI permission model helps you understand the capabilities and safety of the terminal assistant.