Articles from Source: The-New-Stack

AWS’ New Policy Layer in Bedrock AgentCore Makes Sure AI Agents Can’t Give Away the Store

2025-12-02 16:00
AWS has introduced new updates to Amazon Bedrock AgentCore at its re:Invent conference. These enhancements aim to simplify the development and deployment of AI agents on the AWS platform. A key feature is the new Policy layer, which provides security and policy enforcement. This allows businesses to set rules for AI agents, ensuring human validation for certain actions, like offering credits over $100. This addition is part of a broader strategy to enhance trust and safety in AI agents,...
Source: The New Stack
Frederic Lardinois

Give Claude AI Full Access to Your Local Filesystem With MCP

2025-12-02 15:00
🚀 Explore the capabilities of Claude AI with the Model Context Protocol (MCP). This open-source standard enhances AI interactions with APIs, especially in the Claude Desktop environment. Claude Desktop now recognizes files on your system, facilitating smoother workflows. It offers prebuilt connectors, eliminating the need for manual server code, making it more user-friendly for developers. Installation is straightforward, and while the app integrates with the web, users are encouraged to...
Source: The New Stack
David Eastman

Restack Gives Product Teams the Reins To Own AI Agent Behavior

2025-12-02 14:00
Restack is shifting focus to empower product teams in managing AI agent behavior. Co-founders Andrés Tapia and Philippe Brulé recognized a growing demand for AI solutions among their users. As AI tools proliferated, they noticed many companies were launching agents without proper testing. Tapia emphasizes the importance of validating agent behavior after deployment to avoid issues like incorrect information. Restack aims to fill this gap by providing a platform that supports both the...
Source: The New Stack
Meredith Shubel

The 7 Non-Negotiables of AI-Driven Operations

2025-12-02 13:00
In today's fast-paced digital landscape, organizations must adapt to deliver seamless experiences amid rising complexity. AI and automation are essential for enhancing operations and incident response, significantly reducing downtime and mean time to repair (MTTR). By 2025, 67% of enterprise AI investments will focus on integrating AI into core operations. Key to success are seven non-negotiables for AI-driven operations, including true end-to-end incident management and built-in automation....
Source: The New Stack
Cristina Dias

How Okta Scaled From 12 to 1,000 Kubernetes Clusters With Argo CD

2025-12-01 22:00
Okta faced challenges with its Auth0 platform for private cloud customers, prompting a shift to open source GitOps using Argo CD. At KubeCon + CloudNativeCon, engineers Jérémy Albuixech and Kahou Lei shared insights on scaling from 12 to over 1,000 Kubernetes clusters in five years. Argo CD has proven to be an effective solution, supported by a growing community. #Kubernetes #GitOps #Okta #ArgoCD #CloudNative
Source: The New Stack
B. Cameron Gain

Teaching a Billion People to Code: How JupyterLite Is Scaling the Impossible

2025-12-01 20:00
🚀 JupyterLite is transforming education by enabling half a million French high school students to learn math and programming directly in their browsers. This serverless platform, developed by QuantStack, eliminates the need for cloud deployments, making it scalable and efficient. Founder Sylvain Corlay highlights its potential for global accessibility, aiming for a billion users in the future. 🌍💻 #JupyterLite #EducationTech #CodingForAll #TechAccessibility #QuantStack
Source: The New Stack
Michelle Gienow

FlowAI Gives Agents a Greater Role in Infrastructure Automation

2025-12-01 18:01
🚀 Itential has introduced FlowAI, enhancing infrastructure management with reasoning-based agents. These agents can automate processes while ensuring validation and control through the Model Context Protocol (MCP). This innovation aims to balance deterministic operations with AI reasoning, allowing for greater flexibility. FlowAgent Builder enables users to create customized agent personas, integrating large language models for enhanced reasoning capabilities. #InfrastructureAutomation #AI...
Source: The New Stack
Susan Hall

Get Started With NLP Using Hugging Face and Python

2025-12-01 16:00
Unlock the power of Natural Language Processing (NLP) with Hugging Face and Python! 🤖 Hugging Face offers advanced pre-trained models for tasks like summarization, question answering, and sentiment analysis. It outperforms other libraries by utilizing models like GPT and BERT. Real-world applications include market analysis, customer service automation, and research acceleration. 📊💬 A tutorial provides a basic introduction for developers familiar with Python, guiding you through creating a...
Source: The New Stack
Jessica Wachtel

Is Using Command Line AI an Option for Helping You Code?

2025-11-30 19:00
Exploring the potential of command line AI for programming can be intriguing. 🤖 The article discusses using Ollama AI on Pop!_OS Linux to create a Python program. The installation process is straightforward for all operating systems. Once set up, the author tested the AI by generating code for user input and successfully created a functioning GUI app. 📊 This demonstrates the capability of command line AI in coding tasks. #AI #Coding #Ollama #Programming #TechInsights
Source: The New Stack
Jack Wallen

When GenAI Breaks, Rigorous Orchestration Keeps It Running

2025-11-30 18:00
🌐 When moving generative AI (GenAI) to production, success hinges on rigorous orchestration. Only about 5% of pilot projects transition smoothly, often due to mismatched development and production environments. A "shift-left" approach is crucial for early testing and continuous monitoring. Financial institutions face unique challenges, as operational rigor and clarity in roles are essential. Integration across departments is necessary to mitigate risks like unpredictable behavior. Successful...
Source: The New Stack
Manish Gupta

Gaming on Linux? Bazzite Is a Great Place To Start

2025-11-30 15:00
🎮 Exciting news for gamers on Linux! Bazzite is a new distribution designed specifically for gaming. With pre-installed tools like Steam, Proton, and Lutris, it streamlines the gaming experience across various devices. Key features include HDR & VRR support, improved CPU schedulers, and compatibility with multiple gaming controllers. Bazzite also offers security through its immutable design. Check it out for a smoother gaming experience! #LinuxGaming #Bazzite #GameOn #Steam #GamingCommunity
Source: The New Stack
Jack Wallen

2025’s ‘Advent of Code’ Event Chooses Tradition Over AI

2025-11-30 14:00
🎉 The countdown to the 2025 "Advent of Code" event begins! This year marks a significant change as the number of puzzles has been reduced from 25 to 12. Event creator Eric Wastl aims to make the challenge more accessible for participants. Despite advancements in AI, the coding community's enthusiasm for this holiday tradition remains strong. Last year, over 284,000 people engaged in the event, showcasing its popularity. Get ready for another exciting coding adventure! 🧑‍💻🎄 #AdventOfCode...
Source: The New Stack
David Cassel

Broadcom ‘Doubles Down’ on Open Source, Donates Kubernetes Tool to CNCF

2025-11-29 19:00
Broadcom reaffirmed its commitment to open source at KubeCon + CloudNativeCon NA, emphasizing ongoing support for VMware projects. The company announced new contributions to the Cloud Native Computing Foundation (CNCF), including the donation of the etcd-diagnosis tool to enhance Kubernetes reliability. Broadcom also highlighted its involvement with the Cluster API (CAPI) to streamline Kubernetes cluster management. #OpenSource #Kubernetes #CloudNative #CNCF #TechNews 🌐💻🔧
Source: The New Stack
B. Cameron Gain

Python Is Quickly Evolving To Meet Modern Enterprise AI Needs

2025-11-29 18:00
Python continues to evolve, becoming essential for modern enterprise AI needs. 🌐 In 2024, it surpassed JavaScript as the most popular language on GitHub. Its versatility supports various applications, from web services to AI models. However, enterprises must address visibility, performance, governance, and security challenges. Python's growth is fueled by a dedicated community, making it a top choice for AI development. 🚀 #Python #AI #TechTrends #Programming #DataScience
Source: The New Stack
Steve Croce

Most Developers Call AI Data With APIs and A2A

2025-11-29 14:00
🚀 A recent Theory Ventures survey reveals that 91% of senior technical builders are involved in AI development. However, only 17% are building MCP servers. 🔍 The majority, 46%, utilize APIs and A2A for data access. Notably, 72% use AI agents for database access, and 98% assess AI quality through both automated and manual methods. 📊 When it comes to LLM observability, 57% store interactions as traces, while the same percentage use spreadsheets for data review. #AI #TechSurvey #Developers #APIs...
Source: The New Stack
Loraine Lawson

2026 Will Be the Year of Agentic Workloads in Production on Amazon EKS

2025-11-28 21:00
🚀 AWS launched Elastic Kubernetes Service (EKS) in 2018 for early adopters. Now, it's attracting a broader audience who prefer managed solutions. 💡 In a discussion with Mike Stefaniak, key insights on AI workloads and the new Model Context Protocol (MCP) server were shared. The focus is on effective tool naming and a hosted knowledge base to enhance troubleshooting. 🔍 While "agentic AI" is still evolving, it’s expected to gain traction by 2026, moving beyond traditional LLM inferencing. #AWS...
Source: The New Stack
Frederic Lardinois

From Cloud Native To AI Native: Where Are We Going?

2025-11-28 20:00
Panelists at KubeCon + CloudNativeCon discussed the shift from cloud native to AI native technology. Jonathan Bryce emphasized the importance of inference in AI development beyond chatbots. He noted the need for robust models and an inference layer to effectively utilize large datasets. Kate Goldenring highlighted WebAssembly's role in rapid inferencing, emphasizing security through sandboxing. The conversation reflects the evolving landscape of AI technology. #AINative #CloudNative...
Source: The New Stack
Heather Joslyn

Platform Engineering: Where Software-Defined Storage Fits

2025-11-28 17:00
🌐 Platform engineering is transforming the tech landscape by evolving from manual systems management to a sophisticated cloud-based ecosystem. This shift allows developers to focus on innovation without the burden of managing complex data infrastructure. Automated, self-service platforms are key, using software-defined storage (SDS) to streamline processes. Efficiency is enhanced as organizations leverage cloud models, reducing operational overhead and promoting agility across teams. This...
Source: The New Stack
Carol Platz

How Amazon Prepares for Black Friday: Predictive Modeling

2025-11-28 16:30
Amazon engineers discussed their strategies for handling Black Friday traffic during a recent keynote at Kubecon + CloudNativeCon NA. 📈 They emphasized that merely reacting to demand isn't enough. Instead, Amazon uses predictive modeling to prepare for peak traffic surges, ensuring a smooth shopping experience for customers. With Black Friday approaching, these insights could benefit other retailers facing similar challenges. 🛒 #Amazon #BlackFriday #Ecommerce #TechInsights #Kubecon
Source: The New Stack
Joab Jackson

CSS-in-JS: The Great Betrayal of Frontend Sanity

2025-11-28 16:00
CSS-in-JS was introduced as a solution to styling challenges, promising modularity and ease. However, many developers now face issues like runtime CSS parsing and unreadable class names. Performance has also become a concern, as runtime styling can degrade load times and increase complexity during debugging. The initial benefits seem overshadowed by hidden costs, especially for users on lower-end devices. #CSSinJS #WebDevelopment #Frontend #Performance #CodingChallenges 🖥️🔧📉
Source: The New Stack
Alexander T. Williams

Shipping Syslog Data to Elasticsearch With Fluent Bit

2025-11-28 15:00
🚀 Discover how to streamline your logging process with Fluent Bit! This article outlines how to set up Fluent Bit as a centralized Syslog server. It collects logs over UDP and forwards them to Elasticsearch for efficient analysis. 📊 Key prerequisites include Docker, Elasticsearch, and a basic understanding of Fluent Bit concepts. The guide provides a detailed demo using Docker Compose to simulate a logging environment. For more details, check the full article! #FluentBit #Syslog...
Source: The New Stack
Sharad Regoti

Migrating VMs to Containers: What To Know Before You Start

2025-11-28 14:00
🌐 Cloud-native enterprises are increasingly migrating virtual machines (VMs) to containers for various benefits, including cost savings and improved IT efficiency. 📖 In the new eBook by Janakiram MSV, "Running Virtual Machines on Kubernetes," readers gain insights on leveraging Kubernetes for all workloads, including legacy applications. 🔍 Key topics include the advantages of Kubernetes, KubeVirt integration, migration strategies, and operational practices for ongoing success. 📥 Download the...
Source: The New Stack
Vicki Walker

Kamera Uses Simulation To Verify Kubernetes Controller Logic

2025-11-28 13:00
🌐 Tim Goodwin, a graduate student from UC Santa Cruz, is exploring the future of Kubernetes as a universal control plane. He developed Kamera, a simulation software that helps manage and debug Kubernetes controllers without needing a cluster. 📊 Kamera captures the behavior of controllers, enhancing their functionality and making it easier for developers. Goodwin presented this innovation at KubeCon + CloudNativeCon NA 2025. Discover how Kubernetes can manage a variety of systems beyond...
Source: The New Stack
Joab Jackson

The New Role of Enterprise Architecture in the AI Era

2025-11-27 18:00
In the AI era, Enterprise Architecture (EA) is evolving. It now serves as a vital link between business strategy and technology execution. Organizations are encouraged to embed EA throughout the development lifecycle, moving beyond traditional frameworks to support agility and adaptability. This shift is crucial as AI demands rapid responses to changing data and business needs. By fostering transparency and encouraging experimentation, companies can harness EA to drive measurable value....
Source: The New Stack
Ankush Dhar

Platform Engineering’s True Value Is in Application Development

2025-11-27 14:00
🌐 Platform engineering plays a crucial role in application development, streamlining workflows and ensuring consistency in the cloud-native landscape. While it helps optimize internal tools, its true value lies in enabling developers to focus on innovation and business logic. The shift to internal developer platforms (IDPs) transforms infrastructure provisioning into a self-service model, reducing delays and friction. IDPs empower developers by providing easy access to resources and tools,...
Source: The New Stack
Massimiliano Bianchessi

Gnome Foundry Is a New ‘IDE in a Box’ for Linux

2025-11-27 13:00
🌟 Gnome Foundry is set to enhance Linux development with its new modular tools, expanding on Gnome Builder's capabilities. 🛠️ While not a new IDE, Foundry brings automatic features to the command line, catering to Linux users who prefer CLI efficiency. 📦 Key tools include various language servers, build systems, and container systems, making it versatile for developers. For more details on usage and features, check out the article! #GnomeFoundry #LinuxDevelopment #IDE #OpenSource...
Source: The New Stack
Jack Wallen

Cursor 2.0 IDE Is Now Supercharged With AI and I’m Impressed

2025-11-26 21:00
🚀 Exciting news for developers! Cursor IDE has rolled out version 2.0, featuring advanced AI integration. This update introduces the Composer, which is four times faster than similar models, enhancing coding efficiency. Users can now run multiple agents in parallel and benefit from a new interface designed for easier navigation. Additional features include integrated Chrome DevTools, speech-to-text functionality, and improved support for coding tasks. Check it out for a streamlined coding...
Source: The New Stack
Jack Wallen

KServe Joins CNCF To Standardize AI Model Serving on Kubernetes

2025-11-26 20:00
🚀 KServe has officially joined the Cloud Native Computing Foundation (CNCF) as an incubating project at KubeCon+CloudNativeCon North America. This move highlights KServe’s role in standardizing AI model serving on Kubernetes, providing a scalable platform for enterprises. It aims to address the complexities of modern AI workloads. KServe focuses on multinode inference for large language models, enhancing efficiency in cloud-native AI infrastructure. Developed by industry leaders like Google...
Source: The New Stack
Joab Jackson

Beyond Scripts: The Shift From Automation to Agentic AI

2025-11-26 18:00
The landscape of industry engineering is evolving. Over the last decade, automation has streamlined processes from integration to deployment. However, traditional automation struggles with complex scenarios like traffic spikes and unexpected failures. Enter agentic AI. Unlike static scripts, these systems can assess situations and make decisions autonomously. They adapt in real-time, maintaining operational efficiency without solely relying on human oversight. 🌐🤖 Recent studies show that...
Source: The New Stack
Ankush Dhar

Deploy Agentic AI Workflows With Kubernetes and Terraform

2025-11-26 17:00
🚀 AI agents are advancing into multiagent systems with reasoning and memory capabilities. However, deployment remains a challenge for many teams. Kubernetes and Terraform offer solutions for reliable, scalable, and automated deployment of AI workflows. Together, they enable the creation of cloud-native AI systems. The article outlines a process to build and deploy a Python-based AI agent using LangChain, Docker, and Kubernetes. This includes setting up an agentic workflow, containerizing the...
Source: The New Stack
Oladimeji Sowole

SolidJS Creator on Fine-Grained Reactivity as Next Frontier

2025-11-26 16:00
🚀 SolidJS creator Ryan Carniato recently discussed the future of JavaScript frameworks at JSNation. He revealed plans for a new primitive in Solid's reactive architecture. 🔍 Carniato emphasized the significance of signals, comparing them to a spreadsheet for data synchronization. While signals have gained traction, he believes there's more to explore beyond just performance improvements. 💡 His insights suggest that understanding data graphs can unlock new potential for developers. #SolidJS...
Source: The New Stack
Loraine Lawson

3 AI Plays To Survive During Holiday Change Freezes

2025-11-26 15:00
The holiday season is here, bringing change freezes in engineering. These pauses in deployments help reduce disruptions but can also reveal unexpected issues. 🎄🔧 AI technology can enhance monitoring and response during this busy time. It can filter alerts, improve incident visibility, and support teams in managing incidents effectively. 📊🤖 Key strategies include using embedded AI to reduce alert noise and accelerate response by providing relevant context to teams. This ensures engineers can...
Source: The New Stack
Debora Cambe

Tech Veterans’ New Approach To Eliminate ‘Configuration Hell’

2025-11-26 14:00
🚀 Tech veterans are tackling the challenge of "configuration hell" with their startup, ConfigHub. CEO Alexis Richardson highlights the frustration of navigating countless YAML files during system outages. With backgrounds from Weaveworks, Kubernetes, and Salesforce, the team aims to simplify configuration management. Their approach treats configuration as data, addressing misconfiguration issues that lead to significant outages. #TechInnovation #ConfigurationManagement #StartupNews #DevOps...
Source: The New Stack
Susan Hall

How Autonomous Agents Are Changing Infrastructure Management

2025-11-25 20:00
🚀 Infrastructure failures are costly, with downtime averaging $12,900 per minute. AI DevOps engineers are emerging as autonomous agents that analyze infrastructure and coordinate operations in real-time. These systems integrate with tools like Kubernetes and CI/CD platforms while maintaining data ownership for compliance. Key components include local LLM integration and an orchestration layer for managing tasks and approvals. Organizations are adopting six specialized roles for these agents,...
Source: The New Stack
Fahmid Kabir

AWS WAF vs. Google Cloud Armor: A Multicloud Security Showdown

2025-11-25 18:00
In 2025, multicloud strategies are now common, with organizations utilizing AWS, Google Cloud, and Azure. This shift creates challenges for unified app security. AWS WAF and Google Cloud Armor are leading web application firewalls (WAFs) that protect against various threats, including OWASP vulnerabilities, bots, and DDoS attacks. Key differences in integration, rule customization, and automation between the two platforms can impact performance in hybrid environments. Both WAFs play a crucial...
Source: The New Stack
Advait Patel

Amazon CTO Werner Vogels’ Predictions for 2026

2025-11-25 16:30
Amazon's CTO, Werner Vogels, shared key insights at AWS re:Invent, marking a shift in his approach. This year, he provided predictions ahead of the event, hinting it might be his last keynote. His predictions focus on AI and its societal effects, including redefining companionship, quantum-safe encryption, and the fast transition of military tech to consumer products. Vogels emphasized the need for younger voices in tech, reflecting on his 21 years at Amazon. #AWS #WernerVogels...
Source: The New Stack
Frederic Lardinois

How Mozilla’s AI Strategy Disconnects From Its Browser Heritage

2025-11-25 15:37
🚀 Last week, Mark Surman, president of the Mozilla Foundation, shared insights on Mozilla's new AI strategy, indicating a shift from its traditional web focus. 🛠️ Mozilla.ai, an offshoot startup, aims to democratize AI access and promote data ownership. CEO John Dickerson emphasized their mission aligns with the Mozilla manifesto but applies it to AI rather than browsers. 🖥️ Notably, their product line features open source tools, including the recently launched any-llm interface. However,...
Source: The New Stack
Richard MacManus

Self-Service Analytics Failed. Can Agentic AI Finally Succeed?

2025-11-25 13:05
For over a decade, self-service analytics has struggled to meet its promise of universal data access. Key challenges include access control, usability, and semantic understanding. Organizations face difficulties in providing secure data access, navigating complex tools, and aligning on business definitions. Agentic AI, particularly large language models (LLMs), may address these issues. LLMs can simplify SQL queries and enhance context comprehension, helping users interact with data more...
Source: The New Stack
Alasdair Brown

Inside Uber’s Multicloud AI Reality: The Gap Between Data and Compute

2025-11-24 22:00
Uber is transitioning from on-premises to a multicloud infrastructure, highlighting the challenges of optimizing GPU usage across multiple providers. In a recent presentation, Andrew Leung shared insights on how Uber uses AI for various applications like ETAs and fraud detection. However, separating data and compute complicates their infrastructure, requiring careful workload management. The company aims to bridge the cloud divide while addressing GPU scarcity. #Uber #Multicloud #AI...
Source: The New Stack
Alex Williams

Anthropic’s New Claude Opus 4.5 Reclaims the Coding Crown

2025-11-24 19:00
🚀 Anthropic has launched Claude Opus 4.5, its most advanced model yet, excelling in coding tasks with an impressive accuracy score of 80.9%. This release is not only smarter but also more affordable, with reduced API pricing. Users can expect enhanced output in documents, spreadsheets, and presentations. 📊 Opus 4.5 also showcases a new "effort" parameter, allowing developers to optimize resource usage. Notably, it scored higher than previous job candidates in performance tests at Anthropic....
Source: The New Stack
Frederic Lardinois