AI News

Follow the latest AI and automation news

23 Jun 20268 Muharram 1448 AH

Cursor announces its own AI model, a new Git platform, and a mobile app

Cursor has revealed new details about its first AI model trained entirely in-house, marking a significant milestone for the company. Alongside this, it announced a new Git management platform and a dedicated mobile app. This new model is a strategic move to enhance Cursor's competitive edge in the AI market, providing developers with advanced tools for code management. The new platform aims to facilitate collaboration among teams and improve workflow efficiency.

Business

23 Jun 20268 Muharram 1448 AH

OpenAI says new GPT-5.5-Cyber outperforms Anthropic's Mythos on cybersecurity benchmark

OpenAI has announced an expansion of its Daybreak cybersecurity initiative with the launch of updates, including the new GPT-5.5-Cyber model. Additionally, the Codex Security extension has been developed in collaboration with over 25 security companies and multiple governments, demonstrating OpenAI's commitment to enhancing cybersecurity. The focus of this new initiative shifts from merely identifying vulnerabilities to automatically correcting them, reflecting a strategic move towards improving security rather than just recognizing risks.

Governance

23 Jun 20268 Muharram 1448 AH

Warning of AI Attacks on Cybersecurity Readiness in Months

The Five Eyes intelligence alliance warns that AI models capable of launching cyberattacks will be ready in months. They urge governments and businesses to act now to strengthen defenses against these evolving threats. This warning follows a directive from the Trump administration for Anthropic to suspend the use of its advanced models by foreigners, highlighting growing concerns about AI capabilities. Intelligence agencies noted these models could drastically change offensive and defensive cyber capabilities.

💡

Tunisia's RoboCare secures six-figure investment from 216 Capital

What impact could smart agriculture have on regional food security?

Infrastructure

NVIDIA Powers Over 400 of the World’s 500 Fastest Supercomputers

23 Jun 20268 Muharram 1448 AH

Infrastructure

23 Jun 20268 Muharram 1448 AH

NVIDIA Powers Over 400 of the World’s 500 Fastest Supercomputers

NVIDIA announced that its technology powers 81% of the world's top 500 supercomputers, with 26 new systems utilizing the Grace CPU. Data reveals that 90% of the new entries in the list employ NVIDIA technologies, indicating a strong preference for machines designed for AI and simulation. In the Green500 list, the KAIROS system ranks first, featuring the NVIDIA Grace Hopper chip. These figures confirm that accelerated computing has become essential for handling the most demanding tasks in AI and scientific fields, with NVIDIA systems offering more than double the AI training capability compared to other platforms.

Governance

Guterres Urges AI Companies to Disclose Environmental Costs

23 Jun 20268 Muharram 1448 AH

Governance

23 Jun 20268 Muharram 1448 AH

Guterres Urges AI Companies to Disclose Environmental Costs

UN Secretary-General Antonio Guterres has called on major AI companies to disclose the environmental costs of their data centers, emphasizing the need for renewable energy use. This call comes amid increasing scrutiny of data centers due to their high energy and water consumption. Guterres warned that by 2030, data centers could consume more electricity than all but five countries worldwide.

Models

Baidu Unveils Unlimited-OCR: Constant KV Cache Delivers SOTA Performance on Long Documents

23 Jun 20268 Muharram 1448 AH

Models

23 Jun 20268 Muharram 1448 AH

Baidu Unveils Unlimited-OCR: Constant KV Cache Delivers SOTA Performance on Long Documents

Baidu has launched Unlimited-OCR, addressing a key issue in long-document transcription. The model employs Reference Sliding Window Attention (R-SWA) to compress KV cache from linear to constant growth, achieving state-of-the-art results on the OmniDocBench benchmark. Built on DeepSeek-OCR's DeepEncoder architecture, Unlimited-OCR scored 93.92% on OmniDocBench v1.6, surpassing DeepSeek-OCR's 87.01%. This improvement highlights significant advancements in transcription accuracy and efficiency.

Business

ByteDance's Doubao 2.1 Pro Crosses Production Threshold, Seedance 2.5 Video Model Coming July

23 Jun 20268 Muharram 1448 AH

Business

23 Jun 20268 Muharram 1448 AH

ByteDance's Doubao 2.1 Pro Crosses Production Threshold, Seedance 2.5 Video Model Coming July

ByteDance has launched the Doubao 2.1 Pro, surpassing the production capability threshold. This model shows significant advancements in coding, agent capabilities, and visual language understanding, achieving top results in various benchmark evaluations. Doubao 2.1 Pro excels in code generation and multimodal tasks, outperforming Claude Opus 4.6. Its pricing strategy, at 6 yuan per million input tokens, positions it competitively in the market, appealing to cost-conscious users.

Agents

NVIDIA unveils new robot safety platform

23 Jun 20268 Muharram 1448 AH

Agents

23 Jun 20268 Muharram 1448 AH

NVIDIA unveils new robot safety platform

NVIDIA has recently launched a comprehensive safety platform, the first of its kind in the industry, aimed at accelerating the deployment of human-like robots in industrial environments. The new Halos platform integrates smart computing, software, sensors, and inspection tools into a unified safety framework, enabling robotics companies to develop reliable autonomous systems. The platform features the industrial computing platform IGX Thor, the Holoscan sensor bridge, and the Halos OS operating system.

Agents

Nokia to integrate Gemini AI models

23 Jun 20268 Muharram 1448 AH

Agents

23 Jun 20268 Muharram 1448 AH

Nokia to integrate Gemini AI models

Nokia announced its plan to integrate Gemini AI models into its assurance platform in collaboration with Google Cloud. These models aim to enhance the speed of problem resolution for network operators, with six agents being created to deliver significant benefits to users of Nokia's assurance center. The partnership seeks to address challenges arising from the increasing complexity of data generated by networks, which leads to issues in systems reliant on manual troubleshooting. According to the joint statement, the agents will assist in processing raw data instantly, enabling communication providers to identify critical issues.