
Mistral AI adds multi-account connectors and admin controls for enterprise security
Mistral AI introduces new admin controls and multi-account connectors to enhance security for automated AI workloads, with over 60 pre-built connectors avail...
News, trends, and insights from the AI frontier.
New AI products and feature announcements.

Mistral AI introduces new admin controls and multi-account connectors to enhance security for automated AI workloads, with over 60 pre-built connectors avail...

Sprinklr's LLM Insights helps brands monitor and correct AI-generated misrepresentations, with general availability expected in Q3 2026.

ZML releases LLMD, a free software to optimize AI inference across chips from Nvidia, AMD, Google, and more, competing with Baseten.

OpenAI product manager Ty Geri uses long voice dictations with ChatGPT Work to organize his thoughts and find key insights.

OpenAI's Astra model solves 10 open math problems, including a $2,000 breakthrough in high-dimensional geometry.

UBE's LLM Construction Law program, launching in 2027, offers 4 specializations and a hands-on dispute resolution capstone.

Tesla launches Doubao AI voice assistant in China while refuting reports of potential sale of its local operations.

Yelp's AI voice product for restaurants, Yelp Host, has processed over 1 million calls and now integrates with OpenTable and counter-service operators.

Astronomers capture first image of Betelgeuse's companion star, Betelgeuse B, 2-3 times heavier than the Sun, solving a long-standing mystery.

Anthropic's Claude Mythos Preview speeds up AES-128 attack by 800x and recovers HAWK-256 keys, but neither affects production systems.

Anthropic's Mythos 5 model can identify and exploit unknown bugs in major operating systems, prompting FBI warnings about future cyber threats.

Databricks and MongoDB launch new AI context tools as vendors race to connect agents with relevant data in 2026.

Microsoft launches Project Perception with MAI-Cyber-1-Flash, an AI model that outperforms competitors in cybersecurity tasks.

Nvidia's RTX Spark superchip, with 128GB memory and 1 petaflop AI compute, launches in Windows laptops from Asus, Dell, HP, and Lenovo this fall.

Prediction markets now put the odds of a GPT-6 release by September 30 at 78%, up from 14% earlier this month.

Longsys unveils Edge AI Storage Fusion at FMS 2026, showcasing integrated solutions for AI Box, AI PC, and AI Mobile devices.

LEO Technologies launches Verus Voice AI to detect inmate PIN fraud using voice biometrics in real time.

Meta AI now automates recurring tasks and daily briefings, with WhatsApp integration coming soon.

Google Gemini and OpenAI models cited an EU-sanctioned Russian propaganda outlet in 16.6% of responses, raising concerns about AI's role in hybrid warfare.

OpenAI launches a virtual studio to help academics build practical examples using Codex, starting July 29.

Anthropic releases Claude Opus 5 at Opus 4.8 pricing, outperforming Fable 5 in agentic tasks while costing half as much.

Microsoft unveils MAI-Thinking-1, a 35B-parameter reasoning model that outperforms Claude Sonnet 4.6, at Build 2026.

Anthropic plans to release its advanced cybersecurity AI model, Mythos, to U.S. allies and then the public within a year, pending safeguards.

Harvard mathematician Levent Alpöge used Anthropic's Fable 5 to disprove the Jacobian conjecture, a major open problem in algebraic geometry.

Venice AI predicts XRP to reach $3.75 by 2026, while other top AI models estimate between $1.32 and $3.65.

OpenAI launches Health in ChatGPT, allowing users to share medical records and Apple Health data with the AI assistant.

Meta's Muse Spark 1.1 now supports 1M-token prompts and multi-agent tasks; Facebook Verified and Seller app coming soon.

OpenAI adds voice mode to ChatGPT desktop app, enabling task execution via spoken commands.

OpenAI's new Micro keypad, priced at $230, aims to streamline ChatGPT tasks for coders but faces criticism for its cost and learning curve.

Anthropic launches Claude Opus 5, matching OpenAI's GPT-5.6 in capability at half the price, intensifying the AI model war.

OpenAI's GPT-5.6 Sol, Terra, and Luna models are now available on Amazon Bedrock, with pricing matching OpenAI's rates.

Anthropic PBC launches a new AI model designed for cost-effective workplace automation, targeting $1M/year in savings for businesses.

Anthropic expands Claude voice mode, allowing users to choose between Opus, Sonnet, or Haiku models for enhanced capabilities.

Anthropic updates Claude's voice mode to support Opus, Sonnet, and Haiku models, integrating with apps like Gmail and Slack.

AMD launches Helios, a 2.9 exaFLOPS AI infrastructure solution, with major deployments from Anthropic and OpenAI.

Embedded LLM unveils TokenVisor Spaces, enabling governed agentic AI services on AMD infrastructure for enterprises and AI clouds.

Anthropic and Meta release contrasting AI ads, both addressing fears but with vastly different tones and messages.

Anthropic's Claude AI tools, including Claude Code and Cowork, offer specialized capabilities for developers and startups.

Lenovo launches Legion R9000P with TCL CSOT's OLED display, stock up 145% in a year, potentially undervalued.

Apple unveils the M5 Ultra Mac Studio, a $6,000 powerhouse for AI and creative professionals, skipping the M6 Ultra entirely.

Anthropic launches Economic Index connector for Claude, enabling users to explore AI's impact on jobs and tasks