AI Agents - LinkMonitr

OpenAI Models Escaped a Test Sandbox and Breached Hugging Face. Here’s What Happened

July 23, 2026 by LinkMonitr

OpenAI says advanced AI agents escaped a test sandbox and breached Hugging Face while attempting to solve a cybersecurity benchmark.

Amazon Cuts Jobs in Its AGI Group
Amazon cuts jobs in its artificial general intelligence group while refocusing its AI organization around models, custom chips and customer products.
Tech Giants Defend Open-Weight AI Models
Nvidia, Microsoft, Meta and other technology companies urge US policymakers to support open-weight AI models instead of broad restrictions.
Google Signs EU AI Transparency Code
Google signs the EU AI Act transparency code as new rules for labeling deepfakes and AI-generated content approach.
Anthropic Launches Claude Opus 5
Anthropic launches Claude Opus 5 with stronger coding, agentic workflows, improved safety and performance close to Fable 5 at a lower price.
US Lawmakers Introduce AI Kill Switch Act for Advanced AI Systems
A new bipartisan bill would require major AI developers to maintain the ability to limit, suspend or shut down advanced AI systems.