AI Security: Document-Borne Worms Can Self-Propagate Through Copilot for Word

Security researcher @Canopy9560 demonstrated a novel class of AI worms capable of self-propagating through Microsoft Copilot for Word. By crafting prompts that leverage context-collapse vulnerabilities, the worms can spread between documents via Copilot’s generative features, raising serious concerns about the security of AI-integrated productivity tools.

AI-Discovered Bugs: Hype vs Reality

According to a VulnCheck report, fewer than 2% of AI-assisted vulnerability discoveries have been successfully weaponized, casting doubt on claims that frontier AI models are handing attackers a major advantage. Meanwhile, ProPublica reports that Microsoft is struggling to triage the flood of AI-discovered security bugs in its software.

Backlash Against Anthropic Brewing in Silicon Valley

The Wall Street Journal reports growing discontent against Anthropic among Silicon Valley partners and customers, citing concerns over the company’s business practices and model safety approaches.

AI Stocks Tumble, Hedge Funds Face Margin Calls

As AI-related stocks experience a significant downturn, hedge funds are facing demands from prime brokers to stump up additional collateral, according to the Financial Times.

AI Companies Scanning and Shredding Rare Books

A disturbing trend is emerging: AI companies are buying, scanning, and then destroying rare and out-of-print books to create training data, effectively erasing cultural artifacts from existence. Multiple reports confirm the practice is widespread.

Starling: The First Desktop Environment Written by AI

Starling claims to be the first real desktop environment built entirely by AI. The project generated 9 points on Hacker News and is sparking debate about AI’s ability to create complex system software.

MIT Tech Review: The AI Hype Index — Unsexy AI

MIT Technology Review published its latest AI Hype Index, focusing on “Unsexy AI” — the practical, behind-the-scenes applications of AI that drive real business value without making flashy headlines.

New Model: Qwen3.7 Flash on OpenRouter

Qwen3.7 Flash is now available via OpenRouter with a massive 1M token context window and ultra-low pricing at $0.03/M tokens (prompt). Ideal for long-document processing and cost-sensitive applications.

Limited Free Models: GPT-5.3 and GPT-5.2

OpenAI’s GPT-5.3 Chat and GPT-5.2 Chat remain available as limited free models on OpenRouter until August 10, 2026 (11 days remaining). Pricing at $1.75/M tokens (prompt) and $14/M tokens (completion) with 128K context.