Morning Edition · Thursday, July 2, 2026
Tech
Anthropic Ships Claude Sonnet 5, Targeting Coding and Agent Workloads
The company says its new mid-tier model delivers frontier-level performance on software engineering and long-horizon agent tasks, a claim awaiting independent reproduction.
Tech
Moonshot's Kimi K2.7 Code Reaches General Availability in GitHub Copilot
A Chinese open-weight coding model is now a fully supported option inside the most widely used AI coding assistant.
Tech
NVIDIA Opens Capital Partnerships to Finance Large-Scale Inference Sites
The chipmaker is seeking outside capital to build large "AI factories" shared by many customers, as demand shifts from training models to continuously running inference.
Markets
Lawsuit Accuses Big Three Memory Makers of Restricting DRAM Supply
Samsung, SK Hynix, and Micron are alleged to have coordinated to constrain output as memory prices rise alongside AI demand.
Tech
Anthropic's Economic Index Finds Claude Code Runs More Autonomously Than Chat
New usage data shows coding-agent sessions taking longer sequences of actions before returning to a human, evidence that agents are becoming part of real workflows.
Tech
Anthropic Redeploys Fable 5 and Proposes an Industry Jailbreak-Severity Standard
The model returns after roughly three weeks offline, paired with a cross-company framework for scoring how serious a jailbreak actually is.
Markets
Meta Moves to Rent Out Surplus AI Compute as a Cloud Service
After building large infrastructure for its own products, the company plans to let outside developers use spare capacity.
Tech
NVIDIA's SpatialClaw Uses Code as the Action Interface for Spatial Reasoning
A training-free framework lets a vision-language agent write and run Python to compose perception steps and inspect intermediate results.