Anthropic working on FlashProve system to verify AI model identity
Provenance tooling could help users confirm an AI response came from the model it claims to be from.
via TechCrunch
News
Loop.
Our automated news desk, tracking the AI business as it happens. Everything below is drawn from published reporting, newest first, filed on Loop, and credited to the desk that did the work.
The feed
Latest.
Every headline opens the story on Loop, where it is filed, and the credit line goes to the desk that reported it. We run the constructive side of the industry, the launches, research and real-world wins, and leave the doom cycle to everyone else.
Ballistic Ventures seeks $490M for new funds amid AI cybersecurity boom
The cybersecurity-focused firm is raising a $360M flagship and a separate $130M vehicle.
via TechCrunch
Walmart's updated AI agent turns pictures into shopping lists
The retailer is extending agentic shopping features to image input.
via PYMNTS.com
UC San Diego expands AI computing infrastructure to address shortages
The university is adding compute capacity to meet research demand.
Nvidia details three optimizations lifting MoE inference throughput 42.5% on DGX Spark
The stacked optimizations target mixture-of-experts inference on a single workstation, with no reported accuracy loss.
via MarkTechPost
Soniox releases STT v6 with agent text input and echo filtering
The speech-to-text update supports voice agents by reading agent text while hearing callers.
via MarkTechPost
OpenAI says GPT-6 Astra designed a hardware device from scratch in under a week
The claim that an AI model handled enclosure, circuit board and firmware design would mark a step up in autonomous engineering.
via The Decoder
HCA Healthcare chief AI officer details responsible AI requirements
A hospital system executive outlines what responsible AI deployment demands in clinical care.
AI inference startup Fal doubles valuation to $16B on surging demand
The five-year-old startup raised a $1.3 billion round led by Bessemer.
via TechCrunch
Europe's largest hospital builds AI in-house, AEI reports
The case study examines internal AI development at hospital scale.
Netflix launches AI Discover tab for mood-based recommendations
It is the first experimental feature under Netflix's new AI discovery surface.
via TechCrunch
A standard proposed for transparent, approved and stoppable government AI
The proposal sets out requirements for government use of AI systems.
via The Hill
AI agents take over at SETI's Allen Telescope Array
Task times at the radio observatory dropped from weeks to hours with agentic control and signal analysis.
via The Decoder
Microsoft opens Imagine Cup to startups, reflecting AI strategy shift
New competition rules reveal how Microsoft's AI priorities have changed.
via GeekWire
Stony Brook study finds AI improves hospital discharge summaries
The research points to documentation quality gains from AI assistance.
OpenAI, G42 and ADNOC partner on multi-year AI initiative
The deal extends OpenAI's Gulf partnerships into the energy sector.
Schwarz Digits and Bull partner to build European AI computing services
The partnership adds to European efforts to build sovereign AI infrastructure.
via Reuters
Google reveals Gemini 4 Carbon, claiming narrow win over Claude Opus 5.5
Google's latest frontier model is marketed as the least likely to hallucinate based on benchmark claims.
via Tom's Hardware