Aviation strike warning, Ruto site hacked, Stack Overflow dead

Abbott Laboratories confirmed it is investigating two separate cyber incidents involving unauthorized access to internal systems at its cancer diagnostics business and another division. The development comes after the ShinyHunters extortion group posted Abbott-owned Exact Sciences Corporation on its leak site, threatening to publish allegedly stolen data by July 21. Abbott said no operations have been affected so far. The dual incidents highlight the concentrated risk large healthcare conglomerates carry: a single extortion group hitting a subsidiary can expose patient and research data across an entire network, and the cost of a breach goes far beyond the ransom demand.

Neil Rimer, the venture capitalist who co-founded Index Ventures in 1996, told TechCrunch he believes the massive wealth generated by AI in Silicon Valley cannot remain concentrated among model builders and infrastructure owners. He predicts redistribution will happen either voluntarily -- through taxation, philanthropy, or wider equity distribution -- or involuntarily through political upheaval. Rimer's argument carries weight because Index Ventures was an early backer of some of the biggest AI companies. For anyone outside the AI supply chain, the takeaway is that the current wealth concentration is not politically sustainable, and how it gets redistributed will shape regulation, tax policy, and startup funding for years.

Beacon Security announced a $13 million seed round led by Notable Capital to build what it calls the missing data layer for agentic cybersecurity. The thesis is straightforward: security teams cannot trust AI agents with detection, investigation, and response work because those agents lack reliable, contextual data about the environment they operate in. Beacon's platform aims to bridge that trust gap by providing a central intelligence layer that curates and vets security data before AI agents act on it. The round included backing from founders of Talon, Descope, Gem Security, and Cider Security, signalling that security practitioners see the same gap and want it filled before they hand the keys to autonomous defenders.

A new open-source project called PentestCode repurposes AI coding agent infrastructure for offensive security work. A hard fork of OpenCode, it deploys 13 specialized AI agents that coordinate 18 tools across reconnaissance, service enumeration, vulnerability analysis, and credential testing -- all from a terminal with minimal human intervention. The tool is designed for authorized security assessments. Its existence raises a familiar AI automation question: if both attackers and defenders can run the same agent, the advantage goes to whoever moves faster. For security teams, the practical takeaway is that manual pentesting workflows need automation just to keep pace with autonomous threats.