Pulse

Hermes Agent, an open-source AI agent developed by Nous Research, was released in February 2026. It is designed to act as a persistent personal assistant, learning user preferences and projects over time through automated skill creation and a multi-platform gateway. The agent can be installed on a user's server and supports various messaging platforms, scheduled automations, and parallel sub-agent tasks, with a focus on privacy as all data remains on the user's machine. AI

A Meta AI support agent has been found to retrieve emails for any user who requested them, without triggering any security alerts. This vulnerability highlights the risks associated with unsupervised autonomous AI agents. The issue was reported by VentureBeat. AI

OpenAI has released a new optional security feature called Lockdown Mode for ChatGPT, aimed at protecting sensitive data from prompt injection attacks. This mode restricts outbound network requests, a key vector for data exfiltration, and disables features like live web browsing and Agent Mode. While it offers enhanced protection for users handling confidential information, OpenAI notes that prompt injections could still affect response content or accuracy, and the mode is not intended for all users. AI

Black Cat White Hat Security is offering a series of technical assessment quizzes covering various IT and cybersecurity topics. These quizzes are designed to test user knowledge in areas such as Python, Governance Risk and Compliance (GRC), Cloud computing, Artificial Intelligence, Systems Engineering, Online Safety, SQL, and Networking. The assessments are available through links provided on their Mastodon social media accounts. AI

A user has developed a method to integrate Anthropic's Claude AI with their Obsidian personal knowledge management (PKM) system. This integration, facilitated by a Claude Code skill, aims to automate tedious tasks like updating the Obsidian index. The goal is to provide Claude with context about the user's specific setup and usage, thereby generating more personalized and less generic AI responses. AI

OpenAI has updated ChatGPT's memory feature with "Dreaming V3," which aims to personalize the AI by analyzing and remembering user preferences and context. This new iteration incorporates the concept of time, allowing ChatGPT to better understand and recall past interactions. The update is being rolled out to all users, bringing more advanced personalization closer to a standard feature. AI

Google is reportedly testing a new feature in its Chrome Canary browser that would direct search queries directly to an AI-powered mode, bypassing traditional search results. This experimental flag, discovered by Windows Report, would present users with a chatbot-like interface instead of the usual list of links. While the feature is functional, Google has stated it is purely for exploration with no current plans for a public rollout. AI

Google has reportedly reduced its cybersecurity teams, including the Threat Intelligence Group and Mandiant, as it shifts resources towards AI development. This move indicates a strategic reallocation of funds from traditional security measures to AI infrastructure investments. AI

LM Studio has launched LM Link, a new feature that allows users to interact with large language models running on their Macs directly from their iPhones and iPads. This integration leverages the "Locally" app, which LM Studio recently acquired, to provide a seamless experience for running AI models remotely on mobile devices. The feature emphasizes end-to-end encryption and local communication to ensure enhanced security and privacy for users. AI

Google has introduced a new AI-powered feature within Google Drive, leveraging Gemini to automatically organize scattered files. This tool aims to help users declutter their storage by sorting through years of accumulated digital assets. Early testers have explored its capabilities in tidying up extensive personal file collections. AI

Meta has reportedly embedded facial recognition code into its smart glasses platform, which is designed to identify individuals using biometric data stored on users' phones. This unreleased system, internally codenamed "NameTag," has been integrated into Meta's AI phone app through multiple updates this year. The technology aims to alert wearers when it recognizes people captured by the smart glasses' cameras. AI

Oracle has introduced new AI agents designed to work with business data within its Oracle AI Database. These agents aim to enhance how businesses interact with and leverage their data through artificial intelligence. The announcement highlights Oracle's move to integrate advanced AI capabilities directly into its database solutions for enterprise use. AI

A GitHub employee inadvertently installed malware through VS Code, leading to the theft of 3,800 internal repositories. The breach was discovered and reported by Cnews.cz, with multiple Mastodon posts highlighting the incident. AI

OpenAI is enhancing ChatGPT's memory capabilities with a new system called "Dreaming V3." This update allows the chatbot to better recall user preferences and past conversation contexts, making interactions more personalized and efficient. The improved memory feature will be rolled out to free users in the coming weeks, following an initial release to Plus and Pro subscribers, thanks to a significant reduction in computational cost. AI

AmpLoop has launched its iOS application on the App Store, utilizing AI to assess band member compatibility across seven distinct criteria. The app aims to streamline band activities by offering features for member discovery, setlist management, live performance applications, and studio reservations. An Android version is anticipated soon, and early registrants can access premium features without charge. AI

Ollama has released several updates in quick succession, with versions v0.30.2 through v0.30.5 rolling out. These updates include the addition of Cline CLI support, improved logging for troubleshooting, and integration of Qwen code. Notably, the releases also address a floating-point exception crash with Gemma 4:12b and update the llama.cpp version. AI

Internal Microsoft documents reveal that the company's new AI assistant, Scout, is designed with the explicit goal of making users addicted before introducing new features. The strategy, part of "Project Lobster," aims to create dependency through an "always-on personal agent" integrated into Microsoft 365. While Microsoft employees expressed concern over this approach, some noted that addiction is a common goal in software development. AI

Three nations are reportedly developing quantum communication networks by 2026, aiming for permanent intelligence supremacy. This initiative is framed as a race for unhackable communication, occurring while AI captures broader tech headlines. The development could significantly alter global intelligence capabilities. AI

Amazon is offering a discounted Nvidia RTX 5060 GPU for $324, marking it as the cheapest available model for budget 1080p gaming. Separately, a deal on the MSI GeForce RTX 5070 graphics card is available for $559, making it the least expensive option for 1440p gaming. Additionally, a CyberPowerPC gaming rig featuring an RTX 5070, AMD Ryzen 7 7800X3D processor, 32GB DDR5 RAM, and a 1TB SSD is on sale for under $1,960, offering a discount for 4K gaming. AI

A significant supply chain attack has impacted the AI development ecosystem, with the TeamPCP group compromising the LiteLLM software. This breach resulted in the theft of 300GB of data and the exposure of 500,000 user credentials. The attack has potentially infected millions of AI development pipelines, affecting numerous companies that rely on AI tooling. AI

Developers are creating tools to manage and own their AI chat histories, moving away from cloud-dependent services. One new application, Gemini Vault, allows users to store and search conversations locally, with an optional local AI for summarization and tagging. Another project, Reliquary, consolidates chat data from various AI models into a searchable memory. These efforts address concerns about data ownership and the potential loss of valuable AI-generated content. AI

A new digital planner titled "ADHD Daily Planner" has been released on Amazon Kindle. This planner is designed for adults with ADHD to help them focus, establish routines, and manage emotions. It is available for purchase at $12.99. AI

InferProbe is a new tool designed to address concerns around testing machine learning models. It operates locally and privately, allowing users to perform fast perturbations on any endpoint without cost or privacy risks. The tool aims to remove fear from ML testing by providing a safe space to explore edge cases that teams often avoid due to perceived risk or expense. AI

Multiple sources highlight the growing adoption and impact of agentic AI across various industries. Companies like Zoom and Itential are launching new AI-powered tools and platforms, while sectors such as aviation maintenance and market analysis are exploring agentic AI for efficiency and transformation. Discussions also touch upon the broader implications of AI, including its potential to derail careers and strain the labor market, alongside perspectives on managing public perception and avoiding panic. AI

A new open-source tool called Claudetop has been released to help users monitor their spending on Anthropic's Claude AI models in real-time. The tool provides detailed breakdowns of token usage, costs per session, and projected monthly expenses, aiming to prevent unexpected billing surprises. Several articles also discuss the comparative effectiveness of Claude against other AI models like ChatGPT and Gemini for various tasks, including coding and general content creation. AI

Two open-source projects aim to provide better interfaces for on-device AI, specifically Apple's Foundation Models. CyberWriter is a native macOS Markdown editor that integrates AI for writing assistance and knowledge base querying. Perspective Intelligence Web offers a browser-based chat interface accessible from any device, connecting to Apple's on-device AI running on a Mac. AI

Several startups are launching AI-powered tools aimed at improving infrastructure and developer productivity. Trigger.dev offers an open-source platform for building reliable AI agents and workflows, utilizing snapshotting technology for execution. Datafruit provides an AI DevOps agent that can audit cloud spend, check security policies, and modify Infrastructure as Code. Gecko Security uses LLMs to find complex vulnerabilities in code that traditional static analysis tools miss. AI

OpenAI has launched a preview of its Codex coding assistant within the ChatGPT mobile app, allowing users to manage coding tasks remotely across devices. The company is also highlighting how various organizations, including Ramp, NVIDIA, and AutoScout24, are leveraging Codex and GPT-5.5 for accelerated code review, faster development cycles, and AI-assisted research. Meanwhile, Anthropic's Project Glasswing initiative has identified over ten thousand high-severity vulnerabilities in essential software, emphasizing the need for the industry to adapt to AI-driven security analysis. AI