2026-03-31 19:58
Cursor is now allowing Fortune 500 companies to run self-hosted AI coding agents within their own infrastructure. This change addresses concerns over security and compliance, enabling companies to keep their source code and build data secure. 🛡️ Self-hosted agents can access internal tools, dependencies, and networks, enhancing their effectiveness while maintaining tighter security controls. This development comes in response to user demand for more localized solutions. Cursor, founded by MIT...
Source: The New Stack
Paul Sawers
2026-03-31 19:53
🚀 Portkey has launched its open-source Portkey Gateway, designed to manage AI model traffic and enforce policy controls. The move aligns with the growing need for open access in AI infrastructure, allowing teams to operate without multiple SaaS subscriptions. CEO Rohit Agarwal emphasizes that core technology should be accessible for all engineering teams. Portkey already processes 2 trillion tokens daily, highlighting the essential role of gateways in production AI systems. 🌐 #AI #OpenSource...
Source: The New Stack
Adrian Bridgwater
2026-03-31 18:30
🚨 Claude Code users are reporting they are reaching usage limits more quickly than usual. Anthropic confirmed on Reddit that they are aware of the issue and are prioritizing a fix. Users share frustrations, with some stating that a single prompt now consumes a significant portion of their limits. One user noted that a simple greeting costs 2% of their entire session usage. While Anthropic has not identified the root cause, a user claims to have discovered potential bugs that may be inflating...
Source: The New Stack
Meredith Shubel
2026-03-31 16:14
The software industry hit a crucial point in late 2025, as three AI models transformed coding practices. A quarter of Y Combinator’s Winter 2025 startups reported generating 95% of their code with AI, leading to significant productivity gains. 📈 However, the “AI Paradox” arises from fragmentation in tools and processes, limiting AI's effectiveness. Key issues include isolated AI tools, lack of unified data models, and inconsistent trust in AI outputs. Regulatory and budget considerations also...
Source: The New Stack
Manav Khurana
2026-03-31 16:00
AI is transforming application modernization, shifting the focus from maintaining legacy systems to leveraging advanced tools. 🛠️ With AI, teams can scan and update code efficiently, tackling technical debt effectively. However, success still relies on human expertise to ensure accuracy. AI helps simplify complex code, making modernization accessible and less daunting. Combining AI tools with skilled developers is crucial for effective outcomes. #AIModernization #TechDebt #CodeRefactoring...
Source: The New Stack
Matthew Weier O’Phinney
2026-03-31 12:00
🚀 Ollama has released an update leveraging Apple's MLX framework to enhance local AI model performance on Macs. This update aims to improve speed and memory efficiency for developers running large language models (LLMs) directly on their machines. Additionally, support for NVIDIA’s NVFP4 format has been introduced, targeting larger models. The integration allows better responsiveness and reduced latency, making local AI setups more practical for daily development tasks. #Ollama #AppleMLX #AI...
Source: The New Stack
Paul Sawers
2026-03-30 18:45
Kelsey Hightower spoke at KubeCon 2026, emphasizing that everyone is still learning in the AI field. Kristina Kondrashevich from Electrolux shared how Hightower’s insights inspired her team to open-source their project, InfraKitchen. He challenged them during a demo, ultimately approving their work. Discussions at KubeCon centered on the role of AI in open source, highlighting the need for companies to support their dependencies and contribute to projects. #KubeCon2026 #OpenSource #AI...
Source: The New Stack
Jennifer Riggins
2026-03-30 15:00
🚀 Sparky Linux 9.0 has launched, introducing a rolling release model based on Debian! This update brings a variety of desktop environments, including LXQt, MATE, Xfce, KDE Plasma, and more. Users can expect a stable and resource-efficient experience. 🌟 The KDE Plasma version is designed to be minimalistic, maintaining a classic Debian look and feel. Explore the features of Sparky Linux for a fresh take on Debian-based distributions! 🖥️ #SparkyLinux #Linux #Debian #RollingRelease #TechNews
Source: The New Stack
Jack Wallen
2026-03-30 13:00
Microsoft is enhancing its AI strategy by integrating Anthropic’s Claude and OpenAI’s GPT models in Copilot’s Researcher agent. 🌐 Now featuring a ‘critique’ option, GPT drafts content while Claude reviews for accuracy and completeness. This collaboration aims to improve research quality significantly. 📊 Additionally, a new ‘council’ feature allows users to compare responses from different models side-by-side. Microsoft’s Copilot Cowork tool is also being introduced for knowledge workers,...
Source: The New Stack
Frederic Lardinois
2026-03-29 13:00
🚨 Open source software (OSS) is facing challenges as 96% of codebases rely on it, and a surge of low-quality AI-generated submissions is straining maintainers. Many are overwhelmed by verbose changes and nonsensical pull requests, leading some projects to close down. 🎯 Experts warn the traditional contributor model is at risk, with maintainers introducing stricter guidelines. GitHub is exploring solutions to support the community. #OpenSource #AI #SoftwareDevelopment #TechNews #OSS
Source: The New Stack
Bill Doerrfeld
2026-03-29 13:00
WebAssembly (Wasm) is gaining traction as it outperforms containers at the edge. The advancement hinges on finalizing the Component Model, which will enhance its ability to deliver lightweight code with low latency across multiple endpoints. This could position Wasm as a superior choice for edge devices and serverless environments. 🌐⚡ Recent discussions at Wasm I/O highlighted the need for easier implementation and broader support in popular programming languages, aiming for a smoother...
Source: The New Stack
Nick Lucchesi
2026-03-28 20:09
Platform teams are tackling a significant challenge with Kubernetes infrastructure costs. The hidden tax of $43,800 annually arises from managing multiple control planes, with each additional cluster adding to the expense. Virtual clusters like vCluster, Kamaji, and k0smotron are emerging as solutions. They allow teams to provision multiple isolated environments without the overhead of separate control planes, maintaining both cost efficiency and necessary isolation. This shift mirrors the...
Source: The New Stack
Janakiram MSV
2026-03-28 13:30
Nvidia's CEO Jensen Huang discussed the rapid growth of agentic computing at the GPU Technology Conference. 📈 He highlighted a staggering 10,000-fold increase in compute demand over two years. In this landscape, Nvidia's NemoClaw aims to enhance security with three layers, but it doesn't fully address the issues posed by OpenClaw. 🔒 While NemoClaw adds safety measures, it still relies on the open-source model and faces inherent inefficiencies in policy enforcement. The success of such systems...
Source: The New Stack
David Eastman
2026-03-28 13:00
🔍 Understanding data pipelines is crucial for effective AI. A data pipeline is a series of steps that collects, stores, transforms, and delivers data. It plays a vital role in training AI models, shaping their outputs, and enabling them to improve over time. The article also guides readers on building a custom data pipeline and training a model with simulated data inputs. 📊 Learn more about the essentials of data in AI! #DataPipeline #AI #MachineLearning #DataScience
Source: The New Stack
Jessica Wachtel
2026-03-28 13:00
🚀 Solo.io has launched AgentBench to address a key challenge in agentic AI evaluation. This new open-source initiative aims to help developers benchmark autonomous AI systems effectively. With the rise of AI agents, organizations need reliable methods to assess their performance in real-world tasks like infrastructure automation. AgentBench integrates with Solo.io's Gloo Platform, providing a standardized framework for testing AI agents before deployment. It generates detailed logs and...
Source: The New Stack
Steven J. Vaughan-Nichols
2026-03-28 13:00
🚀 Solo.io has launched a new open-source initiative called **agentevals** to address the challenges of evaluating agentic AI systems. This project was announced at KubeCon Europe and aims to help developers benchmark AI agents effectively. With enterprises exploring AI copilots and infrastructure agents, understanding their reliability is crucial. 🛠️ Agentevals provides a framework for testing AI agents in real-world scenarios, integrating with Solo.io’s Gloo Platform and Envoy Proxy. It...
Source: The New Stack
Steven J. Vaughan-Nichols
2026-03-28 11:00
🚀 Anthropic has made headlines in March with over 14 launches and 5 outages, marking a significant period of innovation. Key updates include the release of Claude Opus 4.6, featuring enhanced token capabilities, and Claude Sonnet 4.6, which improves coding and planning tools. The Claude Marketplace also launched, allowing enterprises to use existing funds for new tools without fees. Stay tuned for more developments in AI! 🌐💡 #AI #Anthropic #Claude #Innovation #TechNews
Source: The New Stack
Matthew Burns
2026-03-27 18:00
🚀 The creator of Gitleaks has launched Betterleaks, a new open-source secrets scanner. This tool offers improved detection, faster scanning, and greater flexibility in managing secrets. 🔍 Secrets are crucial for software authentication, but they can be exposed when code is shared unintentionally. Betterleaks aims to mitigate this risk, especially as AI tools increase the amount of code written. 💡 Zach Rice highlights the challenge developers face when relying on AI assistants, which may lead...
Source: The New Stack
Paul Sawers
2026-03-27 17:00
Open source security faces significant challenges following a recent wave of supply chain attacks. 🚨 The Aqua Security Trivy vulnerability scanner was compromised by hackers known as TeamPCP, who infiltrated its CI/CD pipeline. This led to trojanized binaries aimed at stealing sensitive credentials. Since the initial attack, TeamPCP has targeted numerous NPM packages and the LiteLLM Python proxy. They reportedly obtained 300 GB of credentials, raising concerns about their growing impact on...
Source: The New Stack
Steven J. Vaughan-Nichols
2026-03-27 15:20
OpenAI has announced the addition of plugins to Codex, enhancing its functionality with third-party services like Box, Figma, and Slack. These plugins streamline workflows and app integrations, making Codex more appealing to users exploring alternatives. With over 20 plugins available, they support planning and coordination tasks beyond just coding. A dedicated tab in the Codex UI simplifies access to these plugins, with plans for more to come. #OpenAI #Codex #TechUpdates #AI #Plugins
Source: The New Stack
Frederic Lardinois
2026-03-27 13:37
🚨 GitHub has announced plans to use user interaction data from Copilot to train AI models, effective April 24. Users can opt out of this data sharing via account settings. This data includes inputs, outputs, and code context. Notably, Microsoft and other affiliates will have access to this information, but third-party providers will not. Copilot Business and Enterprise users are not affected by this change. #GitHub #AI #Copilot #DataPrivacy #Microsoft
Source: The New Stack
Meredith Shubel
2026-03-27 12:00
Understanding pgvector benchmarks is crucial for successful implementation. 🛠️ While pgvector allows for storing vector embeddings in Postgres, challenges arise when scaling beyond demo scenarios. A recent article highlights the importance of realistic benchmarking, especially when moving from 10,000 vectors to millions. Operational issues can surface at scale, impacting performance. Planning and understanding these factors can lead to better outcomes. 📈 #pgvector #Postgres #DataEngineering...
Source: The New Stack
Naina Ananthaswamy
2026-03-26 15:00
The merger of env zero and CloudQuery highlights a significant issue in the cloud operations market: the Operational Gap. Many enterprises struggle with cloud asset visibility due to inadequate tools, leading to informal management practices. Knowing about resources isn't enough; effective governance and control are essential. This gap, which separates provisioning from ongoing management, is widening, making traditional methods unsustainable. #CloudOperations #TechMerger #OperationalGap...
Source: The New Stack
Yevgeny Pats
2026-03-26 14:00
Enterprise development teams face growing challenges as the speed of code production outpaces validation processes. ⚙️ While coding agents enhance efficiency, they struggle in complex environments with multiple microservices. The existing CI pipelines, designed for slower workflows, cannot keep up with the increased velocity of pull requests. ⏳ Forward-thinking teams are urged to shift validation earlier in the development process to address these issues. 🔄 #SoftwareDevelopment #CIPipelines...
Source: The New Stack
Anirudh Ramanathan
2026-03-25 21:22
Edera, a leading Xen hypervisor company, is set to support KVM this summer, responding to customer demand. Historically, Edera highlighted Xen's security advantages over KVM. However, they now recognize the importance of KVM for organizations with established infrastructures and tooling. Edera plans to adapt its micro-VM isolation model to KVM, ensuring existing systems can run smoothly without needing a complete overhaul. This shift signifies Edera's commitment to meeting the needs of its...
Source: The New Stack
Steven J. Vaughan-Nichols
2026-03-25 15:43
🚀 Kubernetes environments face new challenges as businesses demand AI workloads. Traditional setups struggle with infrastructure drift, leading to unpredictable performance. 🌐 With increasing complexity, especially in regulated industries, teams must move from reactive strategies to a more stable, deterministic infrastructure. 🔧 The solution lies in adopting an API-driven, immutable OS to ensure predictability and security. Join the webinar on April 9 to learn more about scaling Kubernetes...
Source: The New Stack
TNS Staff
2026-03-25 14:39
🚀 Fivetran has announced the donation of its SQLMesh data transformation framework to the Linux Foundation. This open-source tool allows data teams to define, test, and deploy SQL-based transformations efficiently. Key supporters include ATOMS, Benzinga, and Infinite Lambda, among others. SQLMesh, which leverages virtual data environments, aims to enhance performance and offers a strong alternative to dbt. This move follows Fivetran's merger with dbt Labs and reflects a commitment to open-...
Source: The New Stack
Frederic Lardinois
2026-03-25 14:14
HPE is addressing operational fatigue in IT with its new AI agents, designed to assist in root cause analysis. These agents, referred to as "copilots," are being integrated into the OpsRamp platform and aim to reduce analysis time by 50%. This shift marks the beginning of an “Agentic Era” in operations, where AI can take autonomous actions while keeping human oversight. The new application is currently in beta and is set for general release in 2026. As AI complexity grows, many enterprises...
Source: The New Stack
Jennifer Riggins
2026-03-25 13:00
🔍 Online stores often suggest products that don't match user searches, like showing women's boots for "black running shoes for winter." This issue arises from the complexity of modern product discovery systems, which analyze various signals including keyword relevance, shopper behavior, and inventory. Tensor-based ranking can improve this process by evaluating multiple signals simultaneously, allowing for a more accurate representation of product relevance. Traditional ranking methods, while...
Source: The New Stack
Tim Young
2026-03-25 12:00
OpenClaw's rapid rise to fame has also exposed significant security flaws. With over 40,000 instances vulnerable online, issues like data exfiltration have surfaced. 🛡️ In response, Jentic has launched Jentic Mini, a self-hosted solution that enhances safety for developers. It centralizes credentials and offers fine-grained permissions to protect AI agents. 🔐 This initiative aims to fill the gap left by major tech companies. #CyberSecurity #AI #OpenSource #Jentic #TechInnovation
Source: The New Stack
Darryl K. Taft
2026-03-25 11:36
📢 Exciting update from Anthropic! Claude Code and Claude Cowork now have the ability to perform tasks on macOS desktops. Users can ask Claude to open apps, click buttons, and complete various tasks like checking emails or compiling data. This feature works through both direct integrations and traditional screen interactions. Users can also assign tasks via their phone and check the results later on their desktop. Security measures are in place, including permission requests and safeguards...
Source: The New Stack
Meredith Shubel
2026-03-25 11:00
🚀 Building an AI demo is simple, but deploying a reliable AI system in production is challenging. Many projects fail when faced with real-world conditions. Key reasons include: - Poor separation of components - Lack of observability - Cost control issues To succeed, production AI needs proper architecture, monitoring, and scalability. Learn more about the essential patterns and tools for effective AI deployment! 📊🔧 #AI #TechTrends #ProductionAI #MachineLearning #Innovation
Source: The New Stack
Oladimeji Sowole
2026-03-24 16:07
🚀 Exciting news in AI! The Allen Institute of AI (Ai2) has launched MolmoWeb, an open-source visual web agent designed to perform tasks in web browsers. MolmoWeb offers two sizes (4B and 8B parameters) and is built to navigate websites, fill forms, and search for products. It stands out by using synthetic data for training, rather than proprietary models. While it shows strong performance against some benchmarks, proprietary models still lead overall. Ai2 aims to provide an alternative for...
Source: The New Stack
Frederic Lardinois
2026-03-24 16:01
AI-generated code presents security risks, including potentially harmful commands. Dan Phillips discussed this at Wasm I/O in Barcelona, highlighting the need for reliable isolation for AI agents. WebAssembly offers a solution through its unique memory model, avoiding shared kernel pitfalls of traditional sandboxing methods. 💻🔒 For effective execution environments, WebAssembly could be the key. #WebAssembly #AISecurity #TechInnovation #CyberSafety #LLM
Source: The New Stack
B. Cameron Gain
2026-03-24 15:38
Broadcom VMware has announced the donation of Velero to the CNCF sandbox project. This initiative aims to enhance Kubernetes backup and disaster recovery for users across private and public clouds. 🌐 The move responds to user feedback and expands integration with open-source options, providing more flexibility for Kubernetes infrastructure. Broadcom also plans to enhance its vSphere Kubernetes Service with API compatibility for various third-party suppliers. This donation allows the open-...
Source: The New Stack
B. Cameron Gain
2026-03-24 15:20
IBM, Red Hat, and Google have introduced llm‑d, a new Kubernetes blueprint for deploying large language model (LLM) inference, now donated to the CNCF. 🤖 This open-source framework aims to streamline the serving of foundation models, enhancing scalability and cloud-native capabilities. It disaggregates inference processes and incorporates intelligent routing and scheduling to optimize performance. Founding collaborators include NVIDIA and CoreWeave, supporting a community-governed approach. 🌐...
Source: The New Stack
Steven J. Vaughan-Nichols
2026-03-24 14:20
🚀 Brendan Burns, co-founder of Kubernetes, discusses the impact of AI on infrastructure in a recent interview. He emphasizes how AI-generated code will soon be as overlooked as assembly language. 💻 Kubernetes is adapting to AI workloads, requiring a rethink of scheduling and monitoring. GPU scheduling has become crucial, as it involves complex resource requirements beyond just CPU and memory. 🔄 The shift from stateless to batch workloads highlights the need for new strategies in handling...
Source: The New Stack
Frederic Lardinois
2026-03-24 13:00
PwC is reshaping its consulting model by introducing AI agents directly to clients through the new PwC One platform. Clients can now describe their issues, allowing autonomous agents to handle tasks, while consultants review the results. This approach moves away from traditional engagement methods, integrating AI into core workflows. The goal is to create a seamless, self-service experience, with human oversight only where necessary. PwC aims for a more continuous client relationship,...
Source: The New Stack
Darryl K. Taft
2026-03-24 12:43
Discover ExLight, the latest lightweight Linux distribution from Exton Linux! 🌟 Based on Debian 13, it features the Enlightenment 0.27.1 Desktop environment. ExLight is known for its efficiency, stability, and user-friendly interface, making it suitable for both new and experienced users. With tools like Refracta Snapshot and the Calamares Installer, it offers a blend of modern aesthetics and classic functionality. Explore the possibilities with ExLight! 🖥️✨ #Linux #ExLight #OpenSource...
Source: The New Stack
Jack Wallen
2026-03-24 10:00
🚀 Minimus has launched a new initiative aimed at improving security for open-source projects. Eligible maintainers can now access secure container images, SBOM generation, and threat intelligence tools at no cost. This program seeks to enhance software supply chain integrity. 🔒 Minimus emphasizes the need for a ruggedized approach to container security, reducing vulnerabilities and protecting critical digital infrastructure. For developers, this initiative provides essential tools often...
Source: The New Stack
Adrian Bridgwater