Morning Edition · Saturday, June 20, 2026
Geopolitics
US Export Order Cuts Off Access to Two Anthropic Frontier Models
Washington's directive to suspend Fable 5 and Mythos 5 access targets the models themselves, not just chips or downloadable weights.
Tech
DeepSeek Adds a Visual Reasoning Mode to Its Chatbot
The new mode applies chain-of-thought reasoning to images, with the company citing geometry, chart reading, and the conversion of screenshots into HyperText Markup Language (HTML).
Tech
DeepMind Publishes a Control Framework for Advanced AI Agents
The AI Control Roadmap proposes oversight measures that go beyond alignment to constrain agents while they operate.
Tech
OpenAI Reports That Narrow Safety Training Generalized Broadly
Researchers say training a model on realistic human situations improved its behavior on unrelated tasks, an emergent generalization claim that needs replication.
Tech
Figure Says Its Robots Now Outnumber Its Employees
The humanoid-robot company frames the milestone as a shift from prototypes to a deployed fleet, though the comparison is a metric it chose.
Tech
Meta Releases Segment Anything Model 3
The latest version extends Meta's promptable segmentation line further toward open-vocabulary perception across images and video.
Tech
Former White House Adviser Dean Ball to Lead an OpenAI Policy Unit
Ball says he will join OpenAI in early July to head Strategic Futures, a new division focused on AI policy.
Tech
A Single-Image Model Generates Diverse Human Grasps for Any Object
The HUG (Human Universal Grasping) model uses flow matching to produce varied, physically plausible grasps from one color image with depth data (an RGB-D image).
Tech
OpenAI Brings Record-and-Replay Automation to Codex on macOS
The feature records repetitive desktop actions and converts them into a structured, reusable procedure.
Tech
Anthropic Argues Policymaking Must Catch Up to Exponential AI
The lab's policy proposals contend that institutions built for slower change are unprepared for the pace of AI progress.
World
Public Opposition to AI and Data Centers Is Widening
Reports describe resistance moving beyond specific data-center sites to AI more broadly, citing jobs, energy, and trust.
World
Anthropic Launches a National Fellowship for Early-Career AI Workers
Claude Corps aims to extend AI access to communities across the United States through an early-career program.