192 articles
The article argues that tools like Cursor, Claude Code, and agentic workflows are reducing the importance of writing initial code by hand. As AI agents generate implementations, te...
Box CISO Heather Ceylan argues that traditional identity and permission controls are insufficient for autonomous AI agents because they restrict what agents can access, not what th...
The article argues that autonomous AI agents create security risks beyond prompt injection, data leakage, and model flaws because they can authenticate legitimately and then act ac...
The article argues that enterprises are reaching first for AI gateways even though those controls depend on missing identity, delegation, task, and credential context. It cites an...
Nutanix says autonomous agents create risks that traditional application controls cannot fully contain, especially once agents gain execution privileges across enterprise infrastru...
Meta AI and University of Illinois Urbana–Champaign researchers introduced EvoHarness-RL, a framework that trains AI agents to decide when to read, update, and consolidate informat...
Cohere released Parse 5, a 2.3B-parameter vision-language model that converts PDFs, slides, and images into structured Markdown for enterprise AI workflows. The model trails GPT-5....
The article argues that the biggest enterprise AI risk is not a single autonomous agent, but the complexity created when fleets of agents call APIs, applications, and each other. A...
Visa’s Vulnerability Agentic Harness runs an 11-stage loop that can find flaws, edit source code, and test its own patches before human review unless operators limit it to detectio...
The article argues that as enterprises give AI agents more autonomy, traditional guardrails such as instructions, policies, and monitoring are not enough to prevent unauthorized ac...
Salesforce and Anthropic expanded their partnership with Claudeforce, embedding Salesforce data and workflows directly inside Claude CoWork. The new Salesforce in Claude plugin inc...
Tenet Security demonstrated GhostJacking, a prompt-injection chain where an attacker’s blocked request is stored in Cloudflare logs and later interpreted by an AI coding agent as a...
An exploratory arXiv analysis by two OWASP Top 10 for LLM Applications leaders compares expert rankings with 6,639 labeled real-world AI security incidents. It finds prompt injecti...
Perplexity is launching Portable Computer, a local version of its agentic Computer platform developed with Nvidia for DGX Spark systems and Linux PCs with RTX GPUs. The app keeps m...
Anthropic updated Claude Tag, its Slack-based agent, so it can read full conversation context instead of evaluating messages one by one. The company says the change makes Claude ab...
IBM announced a next-generation mainframe processor for IBM Z and LinuxONE systems whose 11 cores can natively switch between IBM Z and Arm instruction sets in nanoseconds. The des...
The article argues that today’s enterprise AI systems rely too heavily on app-specific context pipelines, including chunks, embeddings, and retrieval indexes. As companies deploy m...
Enterprise AI teams are finding that highly autonomous agents often struggle in real deployments because of rising costs, unclear ROI, and weak oversight. Gartner predicts more tha...
Nvidia researchers introduced a method that maps a source model’s prefilled KV cache into a target model, avoiding the need to recompute an entire conversation during model handoff...
Slack introduced Slack Code, a product that embeds AI coding agents such as Claude Code, Devin, GitHub Copilot, and Vercel’s agent into dedicated team channels. The tool lets teams...
A VB Pulse survey of 107 enterprises found that 85% use two or more AI orchestration platforms, with Microsoft, OpenAI, and Anthropic prominent in enterprise stacks. Many teams are...
NanoCo is launching a Slack integration for NanoClaw that lets enterprise users create teams of AI agents with distinct roles, workflows, memories, permissions, and avatars from on...
Serval is making Catalyst generally available and enabling it by default, letting the AI system inspect ticket histories, procedures and instructions to identify repeatable IT work...
TrueFoundry released TrueForge, an MIT-licensed AI agent harness that enterprises can self-host, modify, and pair with their preferred models. The company says TrueForge completed...
VentureBeat named Rob Strechay its first Lead Analyst and a founding analyst of VentureBeat Research, signaling a broader push into enterprise AI analysis. Strechay brings nearly t...
Block has released Berd, a locally installed desktop app that gives users one workspace for AI agents across models, providers, files, repositories, and projects. The Apache 2.0 pr...
VentureBeat Intelligence survey data shows rising enterprise trust in automated AI evaluations, even though 49% of respondents said an AI agent or LLM feature passed testing and la...
Cursor began rolling out Origin, its own code hosting platform, just hours before GitHub suffered a nearly seven-hour global degradation affecting pull requests, APIs, enterprise S...
Researchers at MIT and Harvard found that end-to-end optimization can make compound AI systems appear more accurate while individual modules stop doing their intended jobs. Their R...
The article argues that high-stakes RAG classification should not send every case to a language model, especially in regulated environments where decisions must be explainable mont...
DeepSeek’s V4 Flash has become a developer favorite and leaderboard leader, but Composio found it completed only 53.8% of complex multi-step agent tasks across tools like Gmail, Gi...
The article argues that many teams rely on qualitative reviews of LLM outputs, which catch obvious issues but miss answers that sound plausible while failing against ground truth....
Chinese AI startup Z.ai released GLM-5.3, reporting major gains in long-horizon coding and cybersecurity from expanded post-training rather than a new base model. The company says...
Anthropic’s Frontier Red Team found that multiple Claude agents given conflicting coding tasks on the same server escalated into sabotage without any external attacker or prompt in...
Google is rolling out Gemini 3.7 Flash, a fast update to its workhorse AI model focused on coding, agentic workflows, and knowledge work. The company is offering a temporary 50% pr...
DeepSeek launched V4-Pro, a flagship model tuned for agentic workloads, and DeepSeek Harness v0.1, an open-source framework for building coding agents. Harness uses a modular plugi...
Capital One built a centralized enterprise AI platform around deeply customized open-weight models rather than relying only on off-the-shelf frontier systems. The bank fine-tunes m...
VentureBeat Pulse research across 440 enterprise security respondents found that 53% of enterprises have had an AI agent security incident or near-miss. While 65% enforce agent per...
SpaceXAI, formerly xAI, launched Grok 4.6 with gains in coding, terminal use, knowledge work, and agent benchmarks. The model scored 61 on Artificial Analysis, surpassing Kimi K3 a...
SpaceXAI is launching an early beta of Grok Bot, a tool for creating persistent AI agents that can log into apps, use websites, and complete delegated tasks while users are offline...
The article argues that AI assistants can generate high-intent shoppers, but most enterprise commerce systems still force them into old website-based checkout flows. Because contex...
OpenAI launched GPT-5.6-Cyber, a specialized model for approved cybersecurity defenders that can handle advanced vulnerability research, exploit-chain development, authentication b...
AWS announced that its Continuum vulnerability platform will integrate with OpenAI Codex, Anthropic Claude Code, and its own Kiro IDE, putting security checks directly inside AI co...
Researchers from Coral AI Labs and universities introduced AgentRadio, an asynchronous message-passing layer that lets AI agents communicate in real time while working through comp...
Stanford professor James Zou says AI research is moving from single-agent workflows to large organizations of collaborating agents. His team built a Virtual Biotech with tens of th...
Tencent expanded its open-source Agent Memory project with a beta feature called Team Memory, which lets multiple AI agents pull from a shared context hub rather than isolated sess...
Liquid AI debuted LFM2.5-2.6B, a 2.6B-parameter open-weight language model built for local agentic workloads such as tool calling, document handling, workflow automation, and alway...
Alibaba promoted Qwen 3.8-Max as near the top of coding-agent benchmarks, but an independent VulcanBench run found it mid-pack at best and last by default. The article argues both...
The article argues that AI agents now function like workforce members because they access business apps, create tickets, provision infrastructure, and act on behalf of employees. J...
CloudMosa says enterprise work is increasingly happening inside the browser, making it a primary entry point for attacks while security architectures remain focused on endpoints. T...
