AI issue — AI news, every day at a glance

AI issue is a news media that delivers AI news in an easy-to-read way, every day.

Snowflake Announces 'Cortex AI Gateway,' an AI Agent Management Platform

Snowflake announced 'Cortex AI Gateway,' a new capability for centrally managing AI agent access to enterprise data and tools. The feature can control access from external AI agents—including those from competitors like Anthropic's Claude Code and Cursor—and was accompanied by the announcement of partnerships with five security companies: 1Password, Aembit, Linx Security, SailPoint, and Saviynt. The initiative aims to update traditional security models designed around human users, as organizations enter an era where AI agents operate on behalf of humans.

Snowflake·Jul 29, 2026 09:20 UTC

Latest News

Runway: How a Bug Became a Feature in AI Development

Runway, an AI video generation startup, revealed its model development, evaluation, and release process at VB Transform 2026. The company shared practical development knowledge applicable to projects of any size, including how a real-time video generation bug was converted into a new feature and a simple Excel-based evaluation method.

Runwayml·Jul 29, 2026 05:26 UTC

Runway: How a Bug Became a Feature in AI Development

Instacart Transitions 97% of Engineering Work to AI

Anilban Kundu, CTO of Instacart, a major grocery delivery company, revealed at VB Transform 2026 that in 97% of cases among the company's engineers, AI handles code generation and management, with human engineers rarely needing to read code directly. Through AI utilization, the issue of technical debt has been nearly eliminated, and the accuracy of defect detection in production environments has improved from over 60% to over 90%.

Instacart·Jul 29, 2026 03:21 UTC

Instacart Transitions 97% of Engineering Work to AI

Visa Detects Payment Network Vulnerabilities with AI, Releases Tool Publicly

On June 10, 2025, Visa conducted security verification of its payment network infrastructure using Anthropic's AI model "Claude Mythos" and publicly released the tool "Visa Vulnerability Agentic Harness" on GitHub as open source. The testing was conducted as part of Anthropic's "Project Glasswing," enabling the discovery of vulnerabilities embedded deep within the stack in a cascading manner. The company also published a technical white paper and 12 design principles for critical infrastructure, providing a foundation for other organizations to leverage similar methodologies.

Visa·Jul 28, 2026 23:27 UTC

Visa Detects Payment Network Vulnerabilities with AI, Releases Tool Publicly

Anthropic's AI Discovers Vulnerability in Cryptographic Technology

Anthropic announced that its AI model 'Claude Mythos Preview' discovered a more effective attack method against the post-quantum cryptographic algorithm 'HAWK'. While human experts had spent over two years verifying the security of the algorithm, the AI identified the vulnerability in approximately 60 hours with an API cost of around $100,000. Although there is reportedly no direct impact on currently operational systems, this case is notable as it demonstrates that AI can participate in assessing the security of cryptographic technologies that form the foundation of internet security.

Anthropic·Jul 28, 2026 21:21 UTC

Anthropic's AI Discovers Vulnerability in Cryptographic Technology

GM's Autonomous Driving Division Triples Code Merge Rate with AI Agents

General Motors (GM)'s autonomous driving division has tripled its code merge count by integrating AI agents throughout its engineering workflow, announced VP Rasheid Haque at the 2026 VB Transform Conference. While GM's autonomous driving engineers spend only 15% of their time actually writing code, the remaining 85% is spent on analysis, verification, and troubleshooting—tasks that AI agents have now automated, driving significant productivity gains.

Gm·Jul 28, 2026 19:26 UTC

GM's Autonomous Driving Division Triples Code Merge Rate with AI Agents

AI Enters Lending and Investment Operations in Private Markets

A startup announced plans to introduce AI technology into the private markets sector (investment and lending markets for private companies) to streamline labor-intensive tasks. AI technology is now moving into this field, which has historically relied heavily on extensive human resources for due diligence and lending reviews.

Jul 28, 2026 15:18 UTC

AI Enters Lending and Investment Operations in Private Markets

Neura Robotics to Open Robot Training Facility

German robotics company Neura Robotics announced plans to establish a new artificial intelligence training facility for robots in collaboration with RWTH Aachen University of Technology. The new facility will join the company's already-operating global training data collection network and serve as a hub supporting the development of Physical AI—artificial intelligence for robots operating autonomously in physical space.

Neura-robotics·Jul 28, 2026 09:22 UTC

Neura Robotics to Open Robot Training Facility

Claude's Voice Mode Now Supports AI Model Selection

Anthropic announced improvements to the voice mode of its AI assistant "Claude," which now supports selecting from multiple Claude models and operating them via voice commands. Previously, voice functionality was available only for limited models, but this update expands the selection options, allowing users to choose different models suited to their needs and purposes when using voice control.

Anthropic·Jul 27, 2026 23:20 UTC

Claude's Voice Mode Now Supports AI Model Selection

Moonshot AI Releases Kimi K3 Model

Chinese AI company Moonshot AI has released the model weights and portions of the infrastructure foundation for its large language model 'Kimi K3' as open source. While demonstrating performance levels close to cutting-edge Western models on benchmarks, independent verification has identified significant performance gaps in cybersecurity and mathematics domains, with distillation techniques cited as a potential contributing factor.

Moonshot·Jul 27, 2026 21:22 UTC

European Media Network Announces $1.6 Million Seed Funding

A media network announced on Monday the completion of a $1.6 million seed funding round led by Powerhouse Capital, Axel Springer SE, LadBible, and angel investors from OpenAI and DeepMind. Simultaneously, the company revealed its largest-scale business expansion plan since its founding.

Jul 27, 2026 11:18 UTC

European Media Network Announces $1.6 Million Seed Funding