AI News

Follow the latest AI and automation news

21 May 20265 Dhuʻl-Hijjah 1447 AH

MagenticLite, MagenticBrain, Fara1.5: An agentic experience optimized for small models

Microsoft has launched MagenticLite, an experimental application that operates across browsers and local file systems. It utilizes two small models: MagenticBrain for reasoning and delegation, and Fara1.5 for computer-use tasks. These releases are part of the next generation of Magentic-UI, designed to work together as a cohesive system. MagenticLite reflects Microsoft's vision for agentic AI, enabling the use of smaller models while achieving high performance at lower costs.

Models

21 May 20265 Dhuʻl-Hijjah 1447 AH

The first AI proof worthy of math's top journal landed and it won't be the last

OpenAI's model has successfully disproven a mathematical hypothesis proposed by Paul Erdős in 1946 regarding unit distance geometry. The model employed tools from algebraic number theory, surprising experts in this context. Tim Gowers, a Fields Medalist, regards this achievement as a "milestone in AI mathematics" and warns of the challenges humans face in competing with AI for solving mathematical problems. The disproven hypothesis pertains to the distribution of points in space, asserting that an infinite number of points cannot be placed in a plane at unit distance apart without overlapping.

Infrastructure

21 May 20265 Dhuʻl-Hijjah 1447 AH

NVIDIA GTC Taipei at COMPUTEX: Live Updates on What’s Next in AI

Taiwan is becoming a focal point for AI development as industry leaders gather at NVIDIA GTC Taipei. CEO Jensen Huang will deliver a keynote on June 1 at 11 a.m. Taiwan time, discussing breakthroughs in AI factories and infrastructure. NVIDIA has been recognized at the COMPUTEX 2026 Best Choice Awards for its innovations in AI computing and robotics. The NVIDIA Vera Rubin NVL72 supercomputer won a Golden Award and the Sustainable Tech Special Award, highlighting its advanced capabilities.

💡

White House Issues Executive Order for AI Model Review

How will government review impact AI model development?

Governance

US Cyber Command races to deploy AI on top-secret networks

21 May 20265 Dhuʻl-Hijjah 1447 AH

Governance

21 May 20265 Dhuʻl-Hijjah 1447 AH

US Cyber Command races to deploy AI on top-secret networks

US Cyber Command has initiated a task force to deploy AI models from OpenAI, Google, and others on the most classified Pentagon and NSA networks. This decision follows the demonstration that AI systems like Anthropic's Claude Mythos can identify security vulnerabilities faster than top human hackers. The command aims to leverage these advanced systems, with predictions that similar tools could become widely available within six to 24 months. This move highlights the critical role of AI in enhancing cybersecurity and providing innovative solutions to emerging threats.

Models

Cohere open-sources its strongest model yet

21 May 20265 Dhuʻl-Hijjah 1447 AH

Models

21 May 20265 Dhuʻl-Hijjah 1447 AH

Cohere open-sources its strongest model yet

Cohere has launched its strongest language model to date, Command A+, as open-source code under the Apache 2.0 license. This model marks a significant advancement in artificial intelligence, allowing developers to access advanced technologies without restrictions. This move comes amid increasing competition among major companies in the AI sector. By releasing Command A+, Cohere aims to foster innovation and collaboration among developers, reflecting a trend towards openness in the development of language models.

Business

Anthropic is about to become the first profitable AI lab

21 May 20265 Dhuʻl-Hijjah 1447 AH

Business

21 May 20265 Dhuʻl-Hijjah 1447 AH

Anthropic is about to become the first profitable AI lab

Anthropic is nearing its first profitable quarter, with a projected operating profit of $559 million and $10.9 billion in revenue for Q2, as reported by the Wall Street Journal. Just last summer, the company did not anticipate turning a profit before 2028, highlighting a significant turnaround in its financial performance. This shift is largely driven by demand for coding tools and the use of its agentic model, Claude, which at times exceeded available compute capacity.

Business

OpenAI could file confidential IPO paperwork within days

21 May 20265 Dhuʻl-Hijjah 1447 AH

Business

21 May 20265 Dhuʻl-Hijjah 1447 AH

OpenAI could file confidential IPO paperwork within days

According to the Wall Street Journal, OpenAI is preparing to file confidential paperwork for an IPO within days. This move signals a significant step for the company as it seeks to capitalize on the growing interest in AI. The potential IPO comes at a time when the AI sector is experiencing remarkable growth, highlighting the increasing investor interest in innovative technologies. OpenAI's position as a leader in AI model development enhances its appeal in the market.

Models

DeepMind Launches Genie 3 with Street View Data for Interactive Environments

21 May 20265 Dhuʻl-Hijjah 1447 AH

Models

21 May 20265 Dhuʻl-Hijjah 1447 AH

DeepMind Launches Genie 3 with Street View Data for Interactive Environments

DeepMind has announced the integration of its generative model, Genie 3, with Street View data. This new feature allows users to create AI-supported interactive environments based on real-world locations on maps. Users can select a location on the map and choose a visual theme such as "ocean world" or "desert sands." This technology enhances user experience by merging geographic data with AI, opening new avenues for interaction with digital content. It is part of DeepMind's efforts to improve generative model capabilities and expand their applications across various fields.

Business

SAP taps Mistral AI to help customers migrate legacy software

21 May 20265 Dhuʻl-Hijjah 1447 AH

Business

21 May 20265 Dhuʻl-Hijjah 1447 AH

SAP taps Mistral AI to help customers migrate legacy software

SAP has announced the use of Mistral AI models to facilitate customer migration from legacy software to the S/4HANA system. This initiative aims to enhance customer experience and reduce the time required for the transition. The Mistral AI models are part of SAP's strategy to improve operational efficiency. By integrating these models, customers can leverage AI technologies to enhance performance and lower costs.