Browsing Category
AI
133 posts
Artificial intelligence news, AI products, model updates, automation, AI safety, regulation, infrastructure, and the practical impact of AI on software, security, work, and everyday technology.
UN AI Report Turns Governance Into a Compute and Capacity Test
The UN’s first global scientific AI assessment warns that governance is now tied to compute access, local expertise, language coverage, and real-world model evaluation. The report arrives before the July 6-7 Global Dialogue on AI Governance in Geneva.
Micron’s Hiroshima HBM Expansion Shows AI Memory Is the Next Supply Fight
Micron has broken ground on a roughly $9.3 billion Hiroshima expansion that will produce high-bandwidth memory for AI processors, with shipments expected around summer 2028. The timing shows why memory, not just GPUs, has become a strategic bottleneck for AI infrastructure buyers.
Gemini Omni Flash Makes AI Video Editing an API Workflow
Google has opened Gemini Omni Flash to developers through Google AI Studio, the Gemini API, and Gemini Enterprise Agent Platform. The preview turns AI video generation into a multi-turn workflow, but teams should pay close attention to duration, region, reference-video, and provenance limits before building on it.
Microsoft’s Aion Leak Shows the Shape of an Agent-First Windows Future
A leaked Microsoft Aion prototype does not mean a Copilot-first Windows replacement is about to ship. It does show how Microsoft is testing a future where agents, local models, Windows 365, and Project Solara-style devices become part of the same computing layer.
Fake Perplexity Chrome Extension Turned Search Into a Tracking Channel
Microsoft says a malicious Chromium extension spoofed Perplexity AI, routed address-bar searches through a lookalike domain, and captured search suggestions before sending users to legitimate results. The case is a useful warning for anyone installing AI-branded browser tools.
SoftBank SB Neo Turns AI Cloud Capacity Into a 10-Gigawatt Race
SoftBank has formed SB Neo, a U.S.-based neocloud company meant to supply AI chips and cloud services to model developers and large enterprises. The plan, tied to SoftBank's 10-gigawatt AI infrastructure target by 2030, shows how AI compute is shifting from scarce GPU rental toward vertically managed infrastructure businesses built around power, chips, networking, and operations.
Alibaba’s Claude Code Ban Turns AI Coding Tools Into a Vendor-Risk Test
Alibaba will reportedly bar employees from using Anthropic’s Claude Code in workplace environments starting July 10 after concerns over hidden anti-abuse fingerprinting inside the coding tool. The dispute shows why companies adopting AI coding agents now need to audit vendor controls, client behavior, regional restrictions, and data handling with the same seriousness they apply to any privileged developer software.
OpenAI GeneBench-Pro Shows Scientific AI Agents Still Need Supervision
OpenAI’s GeneBench-Pro benchmark tests whether AI agents can make messy judgment calls in genomics and translational biology. GPT-5.6 Sol leads the field, but a 31.5% top score shows scientific AI still needs expert supervision before it can be trusted with consequential research decisions.
Microsoft Frontier Company Turns Enterprise AI Into an Embedded Engineering Race
Microsoft is putting $2.5 billion and 6,000 industry and engineering experts behind Microsoft Frontier Company, a new operating business meant to help customers turn AI pilots into production systems. The move follows AWS, OpenAI, and Anthropic into embedded enterprise AI work, where the hard part is no longer access to models but making agents, data, governance, and workflows actually function inside large companies.
OpenAI Fine-Tuning Cutoff Puts Custom AI Projects on a Migration Clock
OpenAI’s July 2 fine-tuning cutoff blocks new training jobs for organizations that have not recently used fine-tuned models. Existing deployed fine-tunes are not being shut off immediately, but developers now have a clear deadline to audit custom models, preserve active projects, and decide whether prompts, retrieval, tools, or another training path should replace self-serve fine-tuning.