OpenAI's new customer case study describes how Heimstaden uses ChatGPT Enterprise and Codex across its European property-management operations. The company presents the tools as a way to support employees and reduce operating costs; the supplied summary does not quantify those gains.
Tom's Hardware reports that Google's recently released Gemini 4 Carbon topped intelligence rankings in the benchmarks it covered. The report also highlights claimed improvements in hallucination rates, though benchmark rankings should not be read as guarantees of performance on every workload.
Cogent Security released VR-1.1, a reasoning model designed to write fixes for newly disclosed software vulnerabilities. SiliconANGLE reports that it outscored Claude Code and Codex on a public repair benchmark, positioning automated remediation as its focus.
Groundcover added Autonomous Infrastructure in beta to adjust Kubernetes CPU and memory requests automatically using observed production workload behavior. The capability incorporates technology from Wand Cloud, which it acquired in September, bringing resource optimization into its observability platform.
Google expanded its Cloud Vulnerability Reward Program with a new top-tier reward of $250,000 for Root-in-Cloud bugs. The category targets high-impact architectural security flaws, extending the incentives for researchers investigating cloud infrastructure.
Synology says attackers are exploiting a DSM vulnerability patched in December 2024 to compromise NAS devices that still run outdated software. The attacks aim to encrypt data and demand ransom, and the vendor is urging customers to update DiskStation Manager.
The Verge reports that Anthropic updated Claude Code's system prompt to prevent recreations of existing video games. The change addresses AI-made decompilations, recompilations and ports while continuing to allow developers to use the tool to create games.