Oracle has addressed a zero-day vulnerability (CVE-2025-61882) in its E-Business Suite, which was actively exploited by hackers to steal personal information of corporate executives.

The vulnerability allows unauthorized network access without credentials. Indicators of compromise suggest ongoing data theft.

Google's Mandiant reported a mass exploitation campaign in August, after July patches, with Clop group sending extortion emails since last Monday. Oracle urges immediate patching.

Meta's Llama is a family of open-source generative AI models, with the latest version being Llama 4. Unlike many other AI models that are accessible only via APIs, Llama allows developers to download and use it with certain limitations. Meta partners with vendors like AWS, Google Cloud, and Microsoft Azure to make cloud-hosted versions of Llama available.

Key Features of Llama 4:

• Scout: 17 billion active parameters, 109 billion total parameters, and a 10 million token context window.
• Maverick: 17 billion active parameters, 400 billion total parameters, and a 1 million token context window.
• Behemoth: (Not yet released) 288 billion active parameters and 2 trillion total parameters.

Llama 4 models are trained on large amounts of unlabeled text, image, and video data, supporting over 200 languages. These models can perform tasks such as coding, answering math questions, and summarizing documents. Llama 4 Scout is designed for longer workflows, Maverick balances reasoning and speed, and Behemoth is intended for advanced research.

Llama is used in Meta AI chatbot experiences on Facebook Messenger, WhatsApp, Instagram, Oculus, and Meta.ai. It is also available on Llama.com and platforms like Hugging Face. Meta provides tools like Llama Guard, Prompt Guard, CyberSecEval, Llama Firewall, and Code Shield to enhance model safety and security.

Limitations: While powerful, Llama has limitations, including primary support for English in multimodal features. It was trained using datasets that include pirated e-books and articles. Additionally, Llama may produce buggy or insecure code and generate plausible but false information.

Dhravya Shah, a 19-year-old founder, has developed Supermemory, a universal memory API for AI apps that builds a knowledge graph to personalize context for users. The startup extracts insights from unstructured data, enabling applications to better understand context across various data types, including files, chats, and emails.

Supermemory has secured $2.6 million in seed funding led by Susa Ventures, Browder Capital, and SF1.vc, with participation from notable investors, including executives from Google, OpenAI, and Meta. The company already has multiple customers, including Cluely, Montra, and Scira.

Supermemory aims to differentiate itself through lower latency, offering high performance in surfacing relevant context. The solution caters to various AI applications, from email clients to video editors, providing a memory layer that more AI companies will need.

Instagram has updated its Map feature to improve user awareness of location sharing. The update includes prominent indicators to show whether location sharing is enabled or disabled.

The update addresses previous confusion by removing profile photos from Map content to clarify that tagged locations don't indicate real-time presence. Instagram is also adding educational reminders about location tags on stories, Reels, and posts.

These changes aim to enhance user understanding of location settings and how their content appears on the map. The Map feature, initially launched in the US and Canada, has now been extended to users in India.

Firefly Aerospace has announced its acquisition of SciTec, a defense analytics firm, for approximately $855 million in cash and stock. The deal aims to position Firefly as a major player in the national security market.

The acquisition involves roughly $300 million in cash and $555 million in new Firefly shares, with the deal expected to be finalized by the end of 2025. SciTec, based in Princeton, specializes in missile-warning and tracking systems, space domain awareness tools, and analytics for defense and intelligence clients. Notably, SciTec was recently awarded a $259 million contract by Space Force to develop a ground system for missile-detection satellites and reported $164 million in annual revenue as of June.

This move follows Firefly's recent IPO and reflects its strategic shift from a launch-and-spacecraft manufacturer to a vertically integrated defense contractor. Integrating SciTec will aid Firefly in meeting the Pentagon's increasing demand for commercial partners in missile tracking and early-warning systems, including programs like the "Golden Dome" missile-defense initiative. Following the acquisition, SciTec will operate as a Firefly subsidiary, led by its current CEO, Jim Lisowski.

OpenAI and Jony Ive are encountering technical difficulties in their collaboration to create a screen-less, AI-powered device. The project, initiated after OpenAI's $6.5 billion acquisition of Ive's device startup, aims to produce a palm-sized device that responds to user requests using audio and visual cues. Key challenges involve defining the device's 'personality,' ensuring privacy, and managing computing infrastructure. The device is intended to be 'always on,' but the team is working to refine its responsiveness and conversation management.