2026-03-03 18:17
🚀 Google has launched Gemini 3.1 Flash-Lite, the fastest model in the Gemini 3 lineup. At $0.25/$1.50 per million tokens, it is also the most affordable option. This model is designed for high-volume developer workloads and is available in the Gemini API. Compared to its predecessor, Gemini 2.5 Flash-Lite, it’s more capable but at a higher cost. It performs impressively on multimodal benchmarks, making it a solid choice for data processing tasks. #GoogleAI #Gemini3 #AIModel #TechNews #Innovation
Source: The New Stack
Frederic Lardinois
2026-03-03 17:00
🚀 Google has announced a significant change for its Chrome browser: it will shift to a two-week release cycle starting September 8, 2026, with version 153. This new pace aims to provide users with quicker access to performance improvements and fixes. Currently, Chrome operates on a four-week cycle, which was established in 2021. While individual updates will be smaller to reduce disruption, Google will still offer an extended stable version on an eight-week cycle for enterprise users. Stay...
Source: The New Stack
Frederic Lardinois
2026-03-03 16:38
🚀 OpenClaw has surpassed Linux to become the most-starred project on GitHub, amassing over 250K stars in just four months. This open-source AI agent acts as a personal assistant across various platforms like WhatsApp and Slack. However, questions arise about its long-term viability and potential security risks, especially after a recent hack of its skills marketplace, ClawHub. As interest grows, the community reflects on whether this trend is sustainable or merely a flash in the pan....
Source: The New Stack
Adrian Bridgwater
2026-03-03 15:33
At the Mobile World Congress in Barcelona, discussions focused on how tech companies can leverage AI while adapting to its rapid evolution. Shahid Ahmed from NTT highlighted the telecom sector's potential for reinvention, emphasizing the shift from utility telecoms to AI-driven systems. Key insights included the need for effective governance and the balance between complex orchestration and efficiency in AI deployment. As intelligence moves to the edge, cost implications for GPU usage were...
Source: The New Stack
Adrian Bridgwater
2026-03-03 12:00
Meta is transitioning React into its own foundation under the Linux Foundation, moving away from single corporate control. This change addresses ongoing discussions about governance in open-source projects. Meanwhile, an open letter from advocates opposes Google's new policy requiring direct registration for Android app developers, arguing it centralizes control too much. The balance of community ideals and commercial interests remains a key focus in the open-source landscape. #OpenSource...
Source: The New Stack
Paul Sawers
2026-03-02 21:37
🚀 Anthropic has introduced a memory feature for its Claude chat service, now available to free plan users. This tool allows users to import memories from other AI providers, enabling Claude to understand user preferences right away. Users can edit or disable memory as needed. The app has gained significant traction, becoming the #1 productivity app in the U.S. and seeing a 60% increase in free users this year. 📈 #ClaudeAI #Anthropic #AIMemory #ProductivityApps #TechNews
Source: The New Stack
Frederic Lardinois
2026-03-02 21:30
🚀 Inception Labs recently launched Mercury 2, a large language model that uses a diffusion approach, offering speeds claimed to be 10x faster than models like Claude, ChatGPT, and Gemini. 💡 CEO Stefano Ermon explained that diffusion models refine answers in parallel, contrasting with traditional autoregressive models that generate text one token at a time. 🖥️ Mercury 2's performance is backed by Nvidia’s optimization and is now accessible via an OpenAI-compatible API. For a deeper dive, check...
Source: The New Stack
TNS Staff
2026-03-02 21:00
🚀 CIQ has launched RLC Pro, a new enterprise Linux distribution tailored for AI and high-performance computing. This subscription-based model includes long-term support, compliance packages, and direct engineering support. 🔒 RLC Pro aims to solve the challenges of balancing open-source flexibility with commercial reliability, offering features that are typically add-ons in other platforms. 💼 With bundled support for 3-5 years, FIPS 140-3 validation, and responsive issue resolution, RLC Pro...
Source: The New Stack
Steven J. Vaughan-Nichols
2026-03-02 12:30
Cloudflare has introduced real-time Markdown support for LLM agents, enhancing how AI can interact with web content. This allows for clearer data extraction while optimizing token usage. However, the article raises concerns about the complexity of modern web pages, which often prioritize visual elements over straightforward text. As the web evolves, balancing functionality for both humans and AI remains crucial. 🌐🤖 #Cloudflare #Markdown #AI #WebDevelopment #TechNews
Source: The New Stack
David Eastman
2026-03-01 14:00
🚨 In 2024, a backdoor was found in XZ Utils, a key compression tool in many Linux distributions. This vulnerability could have given hackers control over millions of systems. The issue was discovered by Microsoft engineer Andres Freund, who noticed unusual delays in the SSH protocol. The backdoor was linked to project lead Jia Tan, who took over the project in early 2023. The Commonhaus Foundation, co-founded by Erin Schnabel, aims to support solo maintainers and prevent similar incidents by...
Source: The New Stack
Charles Humble
2026-02-28 15:29
The article discusses the evolution of Physical AI, or embodied AI, as a significant advancement in robotics. This new paradigm focuses on creating autonomous systems that interact with physical environments, moving beyond simple tasks like cleaning. Key developments include self-driving vehicles and collaborative robots in various industries. Experts emphasize the need for advanced foundation models tailored for Physical AI. These models support tasks such as perception and world modeling,...
Source: The New Stack
Bill Doerrfeld
2026-02-28 15:14
🚀 This week in AI developments: Perplexity has launched the Computer, a general-purpose digital worker designed for non-technical users. It automates tasks by connecting with various apps like Gmail and Slack, streamlining workflows effectively. In other news, Anthropic expanded its Claude Cowork with new plugins for popular platforms, enhancing productivity for enterprises. Stay informed and adapt to the evolving AI landscape! 🌐💻 #AI #Technology #Productivity #Innovation #DigitalTransformation
Source: The New Stack
Matthew Burns
2026-02-27 21:00
🚀 Red Hat is advancing its AI capabilities with the introduction of the Red Hat AI Enterprise (RHAE) suite. This new platform integrates AI deployment and management across hybrid environments, built on Red Hat Enterprise Linux and OpenShift. It aims to streamline the AI lifecycle, ensuring consistent and secure operations. Additionally, Red Hat AI 3.3 enhances the platform with new models and tools, including self-service access to privately hosted models. #RedHat #AI #TechInnovation...
Source: The New Stack
Steven J. Vaughan-Nichols
2026-02-26 20:12
Andrej Karpathy has introduced the term “agentic engineering” as a new descriptor for AI-assisted development, moving away from “vibe coding.” This shift reflects the evolution of technology and its practices. Industry voices, including Gene Kim, discuss the challenge of changing established terminology. As developers increasingly adopt AI tools, the need for precise language has grown. Experts predict that this evolution will continue, with significant changes expected by 2026. #AI...
Source: The New Stack
Darryl K. Taft
2026-02-26 18:00
Discover Netdata, a powerful server monitoring tool that offers zero-configuration observability and AI-driven insights! 🌐✨ With per-second metrics available on multiple platforms like Linux, macOS, and Windows, it simplifies monitoring for databases, web servers, and more. Custom dashboards can be created for tailored data viewing. Installation is straightforward with a simple command, and accessing your data is just a browser away! 🔍💻 #ServerMonitoring #Netdata #TechTools #AIInsights #DevOps
Source: The New Stack
Jack Wallen
2026-02-26 18:00
🚨 The surge in agentic workflows is leading to an influx of pull requests, creating validation bottlenecks for engineering teams. Service meshes like Istio can help teams manage this challenge in cloud-native environments. Many organizations are experiencing a disconnect between rapid code generation and the ability to validate it efficiently. Companies like Stripe and Ramp are benefiting from advanced AI workflows, but validation speed is crucial for success. The latest CircleCI report...
Source: The New Stack
Arjun Iyer
2026-02-26 17:05
🚀 Google has launched Nano Banana 2, an upgraded image generation model that focuses on speed and visual quality. This new version, officially named Gemini 3.1 Flash Image, enhances features from its predecessor, including real-time data access and improved text rendering. It can track multiple subjects in workflows and generates better lighting and details. The older Pro model remains available for specialized tasks. #Google #ImageGeneration #NanoBanana2 #AI #TechNews
Source: The New Stack
Frederic Lardinois
2026-02-26 15:28
🚀 Migrating observability platforms can be complex. The article outlines key strategies for using open-source tools like Prometheus, OpenTelemetry, and Fluent Bit. 📋 Start by defining success, identifying services, and determining essential dashboards and alerts for migration. Prioritize what truly matters to ensure a smooth transition. 🔄 Document ownership of services and known quirks to facilitate validation during the process. Open standards can simplify migration, but reaching them may...
Source: The New Stack
Katie Greenley
2026-02-26 13:00
A recent MIT study reveals that 95% of corporate AI initiatives struggle to yield measurable results. The key issue isn’t the AI models themselves but the underlying data architecture. While tools for building AI applications have become more accessible, scaling them for production remains complex due to fragile data foundations. Successful AI deployment requires a unified data architecture to enhance organization, sharing, and real-time delivery of operational data. This approach is vital...
Source: The New Stack
Ajay Khanna
2026-02-26 12:00
The landscape of enterprise IT is shifting. With the rise of Agentic AI, the focus has moved from just containerization to supporting dynamic, resource-intensive workloads. Legacy infrastructure struggles to accommodate these new demands, leading to inefficiencies. Many organizations face high maintenance costs from DIY platforms that drain resources and hinder innovation. As data sovereignty becomes crucial, a new approach is needed: organizations should consume infrastructure rather than...
Source: The New Stack
Oren Penso
2026-02-25 19:00
WebAssembly (Wasm) is a binary instruction format enhancing web app performance, especially for data-intensive tasks. It works alongside JavaScript, which excels in interactivity but not heavy computations. Adopted by major platforms like Figma and Unity, Wasm became a W3C standard in December 2019. Its development addressed JavaScript's limitations in complex applications, evolving from earlier solutions like asm.js and Native Client. Explore how Wasm is shaping the future of web...
Source: The New Stack
Jessica Wachtel
2026-02-25 18:30
VAST Data highlights that trust, not model quality, is the main barrier to AI adoption in enterprises. At the VAST Forward conference, co-founder Jeff Denworth emphasized the importance of trust in AI models and their interactions with data. The company has seen significant growth, surpassing $500 million in annual recurring revenue and launching its first AI Operating System. Their new tools aim to enhance AI workflows for enterprises. #AI #DataTrust #VASTData #TechInnovation #EnterpriseAI 🤖📊✨
Source: The New Stack
Frederic Lardinois
2026-02-25 18:21
Companies investing in open source see significant returns, according to a new report by the Linux Foundation. 📈 Those treating open source as “freeware” face high technical debt and added costs. Nearly 45% maintain private forks, leading to millions in avoidable spending. 💸 The report emphasizes the importance of contributing back to projects for maximizing ROI. Organizations that do often see returns of 2x to 5x their investment. #OpenSource #LinuxFoundation #TechInsights #ROI #TechnicalDebt
Source: The New Stack
Steven J. Vaughan-Nichols
2026-02-25 15:33
Google has introduced an automated review feature for its Conductor AI extension, aimed at enhancing the software development toolkit for developers. This new feature allows for the creation of formal specifications alongside code, ensuring developers maintain control during the coding process. It generates reports on code quality and compliance based on the codebase's specifications. Experts emphasize the importance of automated reviews to prevent issues, comparing it to a chainsaw without...
Source: The New Stack
Adrian Bridgwater
2026-02-25 13:00
In a recent article, the evolving role of databases in the age of AI agents is discussed. Traditionally, databases served human users, focusing on stability and long-term functionality. However, AI agents operate differently, prioritizing speed and responsiveness over ease of use. They work rapidly, often discarding databases once their tasks are complete. This shift necessitates a redesign of databases to cater to AI's fast-paced needs, emphasizing elastic ephemerality and real-time data...
Source: The New Stack
Max Liu
2026-02-24 22:20
📢 Anthropic has launched new connectors and plugins to enhance its Cowork platform for enterprise users. 🔗 Notably, Claude now connects Excel and PowerPoint, enabling seamless data transfer between the two applications. This feature is available for Mac and Windows users on certain plans. 📈 The update also expands the plugin ecosystem, introducing integrations with tools like Google Drive, Gmail, and DocuSign, among others. 💼 Anthropic continues to focus on business use cases, providing tools...
Source: The New Stack
Frederic Lardinois
2026-02-24 22:07
SurrealDB is addressing the challenges AI agents face with traditional enterprise data stacks. 🤖 Current systems often require multiple databases, leading to complexity and operational overhead. SurrealDB's multi-model database aims to simplify this by integrating structured records, connected data, and AI workloads in one engine. Recently, they launched SurrealDB 3.0 and secured $23 million in funding, totaling $44 million raised. 💰 #SurrealDB #AIAgents #DatabaseInnovation #TechNews #Funding
Source: The New Stack
Paul Sawers
2026-02-24 19:00
OpenTelemetry is becoming the standard for organizations needing to instrument applications and standardize telemetry data. 🌐 At the recent OTel Unplugged EU, discussions focused on user needs and improvements, including sampling algorithms and the introduction of a unified observability collector. This collector simplifies monitoring by serving as a single endpoint for all microservices. Experts highlighted the importance of standardizing observability signals to address challenges in...
Source: The New Stack
B. Cameron Gain
2026-02-24 17:00
Many platform teams face a common challenge: they build effective self-service workflows but struggle with user adoption. 🚧 Just because a tool is technically sound doesn't mean it will be embraced. Teams often fail to recognize that internal platforms are products with diverse user needs and adoption curves. 📈 To succeed, platform teams should think like product teams, focusing on user personas and understanding that rollout does not equal adoption. 📚 #PlatformEngineering #UserAdoption...
Source: The New Stack
Oleg Danilyuk
2026-02-24 12:00
🚨 Automation sprawl is affecting engineering velocity in organizations. Many teams rely on outdated Jenkins jobs with minimal documentation, leading to inefficiencies. Hidden costs include time spent troubleshooting and coordinating across platforms. Successful consolidation starts with a maintained inventory of automation tools and accepting that migration takes time. Consider adopting operations-as-code to enhance efficiency. #Automation #Engineering #DevOps #CloudComputing #Efficiency
Source: The New Stack
Justyn Roberts
2026-02-23 23:00
The Pentagon faces challenges with Anthropic's AI, particularly regarding restrictions on using its model, Claude. Secretary of Defense Pete Hegseth met with Anthropic's CEO to discuss these frustrations. While Anthropic aims to loosen some restrictions, concerns about surveillance and weaponization remain. Rob May from NeuroMetric AI suggests enterprises can reduce reliance on single models by using smaller, efficient alternatives. He emphasizes the need for a system that manages multiple AI...
Source: The New Stack
Darryl K. Taft
2026-02-23 19:51
🚀 Amazon Web Services (AWS) has launched Strands Labs, a new GitHub organization for its experimental agentic AI projects. This initiative aims to separate cutting-edge experiments from the stable Strands Agents SDK, which has seen 14 million downloads. The first two projects include AI Functions, enabling code generation from natural language, and Strands Robots, linking language models to physical hardware. While these projects come with documentation, users should anticipate potential...
Source: The New Stack
Frederic Lardinois
2026-02-23 12:00
The article discusses recent advancements in AI coding agents, emphasizing their code-generation capabilities. While many organizations have improved these agents, productivity gains remain limited due to manual workflows and human bottlenecks. Companies like Stripe and OpenAI highlight that the quality of agent output relies heavily on effective feedback loops. A strategy shift towards "harness engineering" can empower agents to self-validate their work, enhancing their autonomy and...
Source: The New Stack
Arjun Iyer
2026-02-22 16:00
Organizations are heavily investing in AI, with some allocating up to 8% of their revenue. However, a recent report reveals that only 1% of companies feel "mature" in AI usage. Despite faster code generation, delivery efficiency is declining. Developers face bottlenecks during deployment, with 74% struggling to scale AI value. Many report their leaders don’t grasp the challenges they encounter. While AI helps with routine tasks, it doesn't address the broader complexities of development. #AI...
Source: The New Stack
Charlotte Fleming
2026-02-22 15:00
🚀 Cloudflare has launched "Markdown for Agents," a new feature that converts web pages from HTML to Markdown on demand. This innovation helps reduce token usage by up to 80%, making interactions with AI agents more efficient. For example, a typical HTML page could weigh 16,180 tokens, while its Markdown version only uses 3,150 tokens. The feature is activated by sending an Accept: text/markdown header, allowing AI tools like Claude Code to benefit immediately. Future updates will enable site...
Source: The New Stack
Steven J. Vaughan-Nichols
2026-02-22 14:00
Unlock the power of WebAssembly (Wasm) for heavy computational tasks in JavaScript! 🚀 This tutorial dives into using Wasm for processing large CSV files, comparing its performance with JavaScript. Key insights include using Rust for better safety and interoperability. Prepare your environment with Node.js, Python, and Rust before starting your CSV processing app. #WebAssembly #JavaScript #Rust #DataProcessing #CodingTutorial
Source: The New Stack
Jessica Wachtel
2026-02-21 16:00
Kubernetes 1.35 introduces significant updates for stateful workloads. The In-Place Pod Resize feature is now generally available, allowing Vertical Pod Autoscaling (VPA) to adjust running pods without eviction. This enhances the management of long-running applications. 🚀 In the article, a tutorial guides you through installing VPA on Minikube, deploying a sample app, and enabling in-place resizing while observing resource adjustments in real time. For effective use, ensure you have Minikube,...
Source: The New Stack
Janakiram MSV
2026-02-21 15:05
🚀 Exciting updates in the AI landscape this week! Roadmap.sh launched the Claude Code roadmap, designed to assist users in navigating AI-assisted coding. This guide covers essential topics for enhancing coding skills and productivity. In an interview, machine learning engineer Stephanie Kirmer shared insights on how LLMs are changing work dynamics, though she warns of inflated investment expectations in the AI sector. A recent study highlights that organizations benefitting from AI are those...
Source: The New Stack
Matthew Burns
2026-02-21 14:00
The recent article discusses the evolving landscape of the Model Context Protocol (MCP) following a conference in London. Developers are addressing challenges in connecting large language models (LLMs) with tools, leading to interest in solutions like "Secure MCP Gateways." Despite its potential, MCP has faced significant hurdles, including security concerns. The distinction between tools and skills in MCP is blurred, raising questions about its future direction. As MCP continues to develop,...
Source: The New Stack
David Eastman
2026-02-20 21:00
🚀 Red Hat has launched the Red Hat build of Podman Desktop, aimed at providing an enterprise solution for container management. This desktop app integrates closely with OpenShift and Red Hat Enterprise Linux, making container workflows more accessible for corporate developers. Podman Desktop offers a user-friendly interface for managing containers, images, and Kubernetes, and is currently in technical preview for qualified customers. With official support, users gain access to security fixes...
Source: The New Stack
Steven J. Vaughan-Nichols