The 10 biggest things in tech, Thu, Jul 9
The ten biggest stories in tech today, scored 1–10 and ranked from 150+ engineering sources. Free, updated every morning.
AI coding assistant symlink vulnerabilities
Wiz researchers disclosed 'GhostApproval,' a critical symlink-based vulnerability affecting six major AI coding assistants — Amazon Q Developer, Claude Code, Augment, Cursor, Google Antigravity, and Windsurf — allowing malicious repositories to execute arbitrary code, while separate research from AI Now Institute revealed 'Friendly Fire' prompt-injection attacks on Claude Code and Codex, and Sophos found these same tools trigger endpoint security rules designed for attackers.
#1 · ai · scored 8/10
AI coding agents security flaws
Multiple research groups disclosed serious vulnerabilities in AI coding assistants this week: Wiz found symlink-based 'GhostApproval' attacks affecting Amazon Q, Claude Code, Cursor, Windsurf, and others; AI Now Institute demonstrated 'Friendly Fire' prompt injection attacks against Claude Code and Codex; and Sophos found these same tools trigger endpoint security alarms through normal behavior, collectively painting a picture of a dangerously under-secured category of developer tooling.
#2 · security · scored 8/10
EU Chat Control advances
The EU Parliament approved Chat Control 1.0 and a related CSAM-scanning proposal narrowly survived a parliamentary vote, advancing binding legislation that mandates automated scanning of private communications and puts direct pressure on encryption standards for all tech companies serving EU users.
#3 · security · scored 8/10
Rust 1.97.0 released
Rust 1.97.0 shipped as a major version milestone for the widely-adopted systems programming language, delivering new features and improvements for practitioners across industry.
#4 · programming · scored 8/10
EU Chat Control advances despite opposition
The EU Parliament approved Chat Control 1.0 and advanced interim CSAM-scanning rules that mandate automated scanning of private messages, imposing sweeping encryption and compliance burdens on messaging platforms and tech companies operating across member states.
#5 · industry · scored 8/10
Historic AI IPO capital concentration
Projected IPOs from Anthropic, OpenAI, and SpaceX are expected to generate more aggregate value than all U.S. VC-backed exits since 2000, signaling an unprecedented concentration of capital in AI that is reshaping tech market structure.
#6 · industry · scored 8/10
Grok 4.5 competitive model launch
xAI launched Grok 4.5 this week, with Elon Musk claiming frontier-class performance matching Anthropic Opus at lower cost, even as a criminal case emerged in which a user generated ~7,000 CSAM images using Grok, raising serious platform accountability questions alongside the competitive positioning.
#7 · ai · scored 7/10
OpenAI GPT-Live voice model ships
OpenAI shipped GPT-Live, a new full-duplex voice model that simultaneously processes speech and generates responses and can delegate complex tasks to GPT-5.5, marking a significant upgrade to ChatGPT's real-time conversational and translation capabilities.
#8 · ai · scored 7/10
Distributed Git for AI-agent era
Former GitHub CEO Thomas Dohmke launched Entire, a distributed Git network designed to replace centralized GitHub hosting and handle the infrastructure strain imposed by AI coding agents generating code at scale.
#9 · ai · scored 7/10
Microsoft Windows zero-days patched
Microsoft patched two significant Windows vulnerabilities in quick succession: RoguePlanet, a zero-day in Microsoft Defender actively exploited in the wild, and CVE-2026-50656, a CVSS 7.8 privilege escalation flaw in the Malware Protection Engine granting SYSTEM privileges.
#10 · security · scored 7/10