AI News Briefing — June 15, 2026

Databricks open-sourced Omnigent under Apache 2.0, a meta-harness that sits above individual agent harnesses like Claude Code, Codex, and Pi to enable composition, governance, and sharing across them. It addresses the fragmentation engineers face when juggling multiple coding and search agents.

A Semafor report cited by The Verge says the White House's export restrictions on Anthropic's Mythos were partly driven by fears that a China-linked group had accessed the model, raising distillation and reverse-engineering risks. The administration has not confirmed the China angle, and a Trump advisor's public statement emphasized other concerns.

Reporting indicates Amazon CEO Andy Jassy and executives from five other companies warned the Trump administration about security vulnerabilities in Anthropic's Fable model, prompting a rapid White House export-control order that forced the model offline. Notably, Amazon is one of Anthropic's largest investors, raising questions about competitive motives behind the security framing.

First spotted on Reddit, now mainstream coverage explores India's sovereign AI debate, Following the suspension of access to Anthropic's newest models, Indian tech leaders are debating whether dependence on foreign frontier labs is a wake-up call for building sovereign AI capacity. The article ties the Anthropic export-control episode to broader questions about India's AI strategy.

A new benchmark called SWE-Explore separates code search from code repair and finds that leading coding agents reliably locate the correct file but miss most of the critical lines within it. The study suggests insufficient context retrieval undermines fixes even when models are otherwise capable.

OpenAI launched a Partner Network backed by a $150M investment to help global partners accelerate enterprise AI deployment and transformation. The program targets scaling adoption through a structured ecosystem of integrators and consultants.

Google Cloud introduced the Open Knowledge Format (OKF), a minimalist spec that standardizes scattered organizational documents into Markdown with YAML frontmatter so AI agents can consume them. It formalizes the LLM Wiki pattern recently popularized by Andrej Karpathy.

KPMG pulled a report promoting AI adoption after it was found to contain fabricated case studies involving organizations like UBS and the NHS. The episode highlights what GPTZero's CEO calls secondary hallucinations, where unverified AI-generated claims spread through trusted consulting channels.