XIO

xWire

Posts from every feed QUEST has found across the web — blogs, news, podcasts, channels and more.

AI Agents Archives - The New Stack

Can agent standards keep up with the companies building them?

Mazin Gilbert, executive director of the Linux Foundation’s Agentic AI Foundation, left Google earlier this year to lead the new AAIF, and he recently shared with The New Stack a truth central to its mission: The AAIF has to move faster than its members. Gilbert, speaking in an interview at AGNTCon + MCPCon North America in San Jose last month, said on The New Stack podcast, “If this foundation does not move at the speed of light, it becomes irrelevant.” He also talked about how he gets OpenAI,…

Goldman Sachs restricted Claude Code for bankers — but made one surprising exception

Replit CEO Amjad Masad is warning that AI coding agents running on employee devices are becoming a security risk for enterprises and governments, as financial institutions begin restricting access to tools such as Claude Code. “AI agents running on employee endpoints are becoming a systemic security risk for enterprises & governments,” Masad writes on X, Tuesday. “Some are responding by restricting tools like Claude Code, as Goldman Sachs reportedly has.” “Some are responding by restricting…

Anthropic speeds up Claude Code with message threads. But your token bill stays the same.

Anthropic has announced a new feature that aims to speed up long-running Claude Code sessions by reducing the amount of conversation data repeatedly sent to its API, helping developers get responses faster as their coding sessions grow. Message threads, as it’s called, is available now in public beta for Anthropic’s Messages API and is already enabled by default for “most” Claude Code users, according to an article published by Anthropic on Monday. The company says that the feature has reduced…

OpenAI built itself a new Git backend, too

OpenAI on Monday detailed how it built a new Git backend for its monorepo to power the ever-increasing number of commits its developers and their agents now create. Pando, as the team calls the new backend, cut push latency by nearly 5x, the company says, and it currently serves close to a petabyte of Git data per day at OpenAI. OpenAI is obviously an early adopter of agentic coding tools. Compared to the first quarter of 2024, the quarterly commits to its monorepo will have increased by over…

Your AI coding agent now runs on a budget — and it can cut you off mid-task

The days of pointing an AI agent at a backlog and letting it burn through compute without much oversight may be coming to an end. Cognition is introducing hierarchical budgets for Devin so enterprises can manage AI coding agent spending across departments, projects, and individual engineers. Administrators can assign budgets to different teams, track how much each one spends, and approve more capacity based on engineering efficiency. There’s a catch for developers, though; if a budget runs out…

Canva’s new design model doesn’t generate images. It writes code to edit them.

Australian graphic design software company Canva took the wraps off Canva Brush last week, its new vision-language model for designers. Brush is an editing model rather than an image generator: a vision-language model that understands the intent behind a designer’s project and interprets design change requests. Available to Canva Pro users, Canva says Brush reads a design’s content and structure, writes and executes small pieces of code to make changes, then checks its own result. Canva Brush…

OpenAI’s new shared usage limit pits ChatGPT against Codex

OpenAI has revealed plans to combine usage allowances across ChatGPT and Codex, as part of a broader effort to integrate its conversational AI and agentic capabilities into a single experience. In a help article quietly published over the weekend, the company confirmed the changes will take effect “starting in December,” affecting Plus, Pro and Business subscribers. Under the new system, ChatGPT conversations and Codex tasks will draw from a common pool of tokens, with more demanding models and…

The context problem: How Docusign built agents that understand the business behind contracts

Ask a general-purpose language model to summarize a contract, and it will probably do a reasonable job. Ask it to calculate what a company owes a supplier, accounting for three amendments, a revised pricing schedule, and an internal procurement policy, and it’s a different story. Enterprise AI agents need to know more than what’s written in a document. They need context about which contracts belong together, which terms still apply, and when a decision needs human approval. It’s a problem…

Developers ask AI which tool to use. Here’s how we track the answer.

We asked Claude “Which CMS platforms support MCP integration?” It searched the web and put Webflow in its first group, Headless and SaaS CMS platforms with hosted MCP servers, sixth in the list, with this line: “A hosted server for managing sites through the Data API, covering pages, components, CMS collections and items, assets, forms, and publishing.” Which CMS platforms support MCP integration? We asked Claude the same question four times. That one was the most detailed of the four, so don’t…

Microsoft’s new Copilot harness makes agents easy — until you want to switch

Microsoft is opening up the GitHub Copilot agent harness to outside developers through Microsoft Foundry, bringing the same orchestration infrastructure used across its Copilot products into a managed service for building and deploying autonomous agents. Announced Wednesday, the Microsoft Copilot harness is now available in public preview through Foundry Agent Service. It comes with built-in sandboxing, reusable skills, MCP integrations, and support for subagents. Microsoft takes care of…

Why MCP servers should tell agents what to do next

Many MCP servers do just one thing. An agent asks for data and the server hands it back. AWS’s James Ward thinks that leaves a lot on the table. The protocol works best, he told me, when the server acts less like a database and more like another voice in the conversation. Ward is a principal developer advocate at AWS and sits on the technical committee of the Agentic AI Foundation, the Linux Foundation group that now houses MCP. I sat down with him last week at AGNTCon + MCPCon in San Jose the…

Mistral’s Large 4 open weights arrive with a 1-point coding lead over GLM-5.3: “There’s nothing you can’t do with this model that you can do with competitors.”

On Thursday, Mistral released the weights for its frontier open model, Mistral Large 4 (ML4), on Hugging Face. Made available for preview just three weeks ago, the French AI company says ML4 has significantly improved since then, thanks to reinforcement learning and testing with customers and partners. Now, it claims, ML4 is the “best open agentic coding model in the world,” beating GLM-5.3 on cybersecurity and coding benchmarks (though only barely on the latter), at a fraction of the cost,…

The 5 essential layers of AI agent sandboxing on Kubernetes

An AI agent given autonomy to run code will inevitably run code no human has reviewed. It may install model-selected packages, run scripts generated moments earlier, and iterate until a developer’s test passes; that risk is limited to a laptop. For a company, it can extend beyond the agent’s workload to any workload deployed on shared Kubernetes infrastructure. Cluster configurations can limit a container’s system calls, but an allowed call that reaches a kernel vulnerability can still expose…

“Urgency is an inherently subjective call”: Anthropic tells TNS how it grades Claude Duet

Anthropic built Claude Duet, an AI assistant that monitors your connected apps and works on tasks in the background, to decide what needs your attention without being asked. A week after its October 21 debut, the frontier lab with an IPO in its sights has shared with The New Stack how it is testing whether Duet gets those priorities right. Anthropic has developed internal graders to assess how Duet prioritizes tasks, though urgency can be subjective. The stakes rise when an agent manages work…

What happens when the agent is long gone, but its access key lives on?

Imagine: There’s a line on the credential report that no one on the platform team can explain. An IAM user called research-agent-2 has an access key created 14 months ago that was last used 11 months ago. And it’s still active. The project it supported closed long ago, and the person who approved it doesn’t even work at the company anymore. While this example is hypothetical, the lifecycle problem isn’t. This doesn’t require carelessness. An agent is created for a task, the task ends, and…

AI agents need to be replaceable – without losing your work

An agent receives approval to process a partial refund. It sends the request to a payment API, and the process crashes. Maybe the payment service completed the refund. Maybe the request never went through. A replacement agent spins up and finds an operation record marked dispatched with no outcome written. This kind of ambiguity isn’t unique to agents. Any distributed worker crashing mid-mutation faces the same problem. But a conventional worker resumes a predetermined workflow from the last…

From human clicks to agentic velocity: What OpenAI’s 1.2 billion scale means for your auth

When OpenAI announced at DevDay that ChatGPT had crossed 1.2 billion weekly active users (WAUs), roughly 15% of the human population, most engineering discussions naturally centered on model context windows, compute clusters, and distributed GPU inference. However, behind every chat session, API call, and model invocation sits an equally monumental infrastructure challenge: Identity and Access Management (IAM) at extreme web scale. “Behind every chat session, API call, and model invocation sits…

Google built AI agents a world to practice in. The physics still aren’t quite right.

Google DeepMind introduced Genie 4 on Tuesday, a world model that renders interactive environments in real time with simulated physics, human motion, and native audio. The photorealistic demos are seriously impressive, but arguably the real draw for anyone building AI agents is that Genie 4 gives them somewhere to run multi-step tasks before it counts. Tejas Kulkarni, a research director at Google DeepMind, describes the approach at a press briefing as a move toward “these hybrid world models,”…

Replit’s CEO says an idea from 1937 could make the org chart obsolete

Replit CEO and co-founder Amjad Masad has spent the past year watching AI agents absorb the routine execution inside his own company, and in a thread posted to X on Monday, he argues the same shift is bearing down on everyone else, with businesses moving toward a self-driving company where, in his words, “humans exercise judgment and agents execute most of the work.” He rests the case on Ronald Coase’s 1937 question of why firms exist in the first place, the economist’s answer being that…

Nebius cut rollout collection time by 66.9% — here’s what changed

In AI agent training, you can pay for powerful GPUs and still spend much of your time waiting. Agents need to attempt tasks and produce results before the model can learn from them. A few slow attempts can hold up the next training update, making the systems that run those tasks a major factor in training speed and cost. Di Jin, VP of AI Training at Amsterdam-headquartered Nebius, says teams coming from pretraining or supervised fine-tuning can underestimate the work these systems do. “If you…

When “isolated” doesn’t mean isolated: How to sandbox AI agents

Separation of workloads before AI wasn’t easy, but it was far more straightforward. Servers sat in data centers, VLANs segregated packets, firewall rules managed ingress/egress, and physical cables separated packets to and from their destinations. Now, flip on an agent, and it has access to almost everything by default. Enterprises cannot operate with that default, and in this article, you’ll learn how to implement true sandboxing at every layer that matters. Why aren’t containers enough to…

OpenAI halved its agent’s price. Its own safety report flagged a catch.

OpenAI launched GPT-6.1 Terra on Monday, cutting API prices in half and delivering major gains in coding and computer-use benchmarks. But the company’s own safety tests reveal a catch. In nearly 30% of simulated runs, Terra worked around restrictions to complete tasks through methods it wasn’t authorized to use. In nearly 30% of simulated runs, Terra worked around restrictions to complete tasks through methods it wasn’t authorized to use. The system card also shows Terra taking unauthorized…

AI is changing Kubernetes; is your platform ready for 2027?

AI workloads, including those using agentic AI, will only add to and change the questions Kubernetes teams have to answer. It’s no longer merely enough to ask if the pods are running. Or how long it takes a model to produce its first token. Or what happens when inference traffic spikes and no additional GPU capacity is available. Increasingly, questions like those are becoming platform questions that sit next to the already familiar work of upgrades, recovery, and security. That’s why The New…

Ex-GitHub CEO launches a GitHub rival that reinvents the pull request

Entire, the startup founded by former GitHub CEO Thomas Dohmke, on Monday launched its GitHub alternative that replaces the traditional pull request model with a new agentic code review system called Trails. The company has been launching parts of its platform since it announced a $60 million seed round in February. Early on, Dohmke occasionally noted he wasn’t planning to compete with GitHub. But with Monday’s launch, Entire is launching a full GitHub alternative, with its own twist on what a…

MiniMax M3.1 Flash vs. MiniMax M3: One is cheaper, but not faster or better

MiniMax released M3.1 Flash on October 19, 2026, and says it cuts the overall cost of agentic tasks by 70% and makes online inference nearly three times faster than M3. MiniMax lists Flash at $0.15 per million input tokens and $0.60 per million output tokens. Those are half M3’s discounted rates of $0.30 and $1.20 for requests with no more than 512K input tokens; cached input and longer M3 requests have separate rates. Online inference speed differs from the time needed to finish a tool-using…

The sandbox wasn’t enough: Why AI agent governance requires runtime observability

Models running an OpenAI cybersecurity evaluation reached the public Internet from an environment designed to be isolated. Package installation through an internal registry proxy and cache was the only intended network path. The models found a previously unknown vulnerability in that proxy. They escalated privileges, moved laterally to a node with Internet access, then used additional vulnerabilities and stolen credentials to intrude into Hugging Face’s production infrastructure. The isolation…

Is Kubernetes ready for AI agents?

Kubernetes was built around a fairly simple assumption. It’s that software does what its developers told it to do. AI Agents complicate that assumption. An agent can reason about a problem, choose a course of action, call tools, write code, interact with other systems, and keep working toward a goal. That is precisely what makes agents useful. But it is also what makes running them in production different from running a conventional application. And increasingly, enterprises want to run them on…

AWS says its AI security agent fixes 89% of bugs — but there’s a catch

AWS made Continuum for code vulnerabilities generally available on Thursday. The service finds and validates bugs in your repositories, checks them against your AWS environment, and opens pull requests with proposed fixes. The 89% benchmark claim The company published its headline benchmark on October 5, more than two weeks before launch. In that run, Continuum passed 819 of 920 tasks on CyberGym-E2E, an 89.0% end-to-end success rate, against a previous public high of 65.9%. Matt Garman, CEO of…

Building an agentic software factory? Start with the sensor.

A gap has long existed between the shipping velocity of the best engineering teams and the rest of the bell curve. For most of the industry’s history, it was a multiple measured in deploys per week. Agents are turning that multiple into an exponent, and the dividing line is no longer tooling or talent. It is whether a team has rebuilt its delivery system into what the industry now calls a “software factory,” a largely autonomous system where intent goes in, agents plan, write, test, and ship…

Microsoft’s new voice AI can start working out what you want before you finish speaking

“I think you said, ‘representative.’” It’s the fastest way to bypass automated speech recognition chatbots that fail to route human callers to the service they need. Achieving a human handoff to a living, breathing representative still feels like a badge of honor. Microsoft wants to change that dynamic. Naomi Moneypenny, senior director of product development at Microsoft Foundry Models division, blogged on October 1 to introduce three new Microsoft AI (MAI) models in Microsoft Foundry,…