AI Agents news and analysis on Web Pulse: 49187 curated articles tracking the latest AI Agents developments, tools, and research, updated continuously from vetted sources.
The U.S. Office of the Comptroller of the Currency issued its first supervisory framework for AI in banking on October 22, 2026, directing examiners to weigh the consequences and reversibility of AI failures rather than …
Ossprey Security obtained and analyzed the full source release of Miasma, a self-propagating credential-stealing worm published on npm as `voicefromtheouterworld` that needs only a single GitHub PAT to harvest credential…
A developer ran three fixed Python CSV-summing programs under Linux RLIMIT_AS caps to show that a MemoryError in an agent evaluation can reflect the runner's resource limits rather than a flawed solution. The bulk-loadin…
Uber built an internal MCP Gateway that now hosts over 800 MCP servers and more than 5,000 tools, centralizing how AI agents reach the company's back-end services, Uber engineers reported. The platform pairs an MCP Regis…
Oracle Corp. has rebuilt its data stack around its Autonomous AI Lakehouse, an open layer that mounts other catalogs such as Databricks' Unity Catalog along with Postgres and MySQL databases so data is queried in place r…
Oracle Corp. is urging database customers to prepare for quantum security risks, with Vipin Samar, Oracle's senior vice president of software engineering for database security, warning that encrypted traffic copied today…
Google Search Product VP Robby Stein said Google is now letting home service businesses receive bookings and quote requests generated in AI Mode through partner platforms Angi, Thumbtack, Housecall Pro and Jobber, with S…
CloudOpsBench v1.0, a new benchmark of 45 real AWS DevOps tasks, found that Claude Fable 5.1 completed 29 of the 45 tasks and GPT-6 Astra completed 26, with roughly 80% of failures traced to reasoning gaps rather than mi…
A paper submitted to arXiv on 3 November 2026 reports that training a draft model to be rejected at critical decision points in speculative decoding raises end-to-end success rates on complex agentic coding benchmarks wi…
The East Corp Capital Open Data & API Desk has published OpenAPI 3.1.0 specifications and Model Context Protocol toolkits that encode 2026 East Coast statutory tax cliffs as piecewise-discontinuous functions for use by L…
A benchmark called AbMaze, built by Dominik Scherm and published November 4, 2026, found that no frontier AI model outperforms a scripted Bayesian optimizer at navigating a fully measured antibody fitness landscape of 65…
Jinna launched as a product for solopreneurs that combines a "multilayer memory" of the user, their business and their clients with AI agents that watch inbox, calendar, tasks and documents and then draft proposals, send…
Nodus published a worked analysis of the compute required to run reinforcement-learning training for LLM agents, showing that GPU cost is driven by idle time while tool calls execute in CPU sandboxes. In its example, 128…
A developer building a LangChain-based survey-report agent found that Anthropic's Claude Haiku 5.5 generated 17 bad citations in total by truncating 36-character respondent UUIDs to their first 8 characters, a failure th…
Ricursive Intelligence released Ripple-65B Preview, a 65B-parameter open-weight model it says reaches U.S. open-weight state of the art on agentic coding with an 80.52 score on TerminalBench 2.1, ahead of Reflection Beam…
Enrich Labs launched Agent Conversion Optimization on November 4th, 2026, a product that tracks whether personal shopping agents recommend an e-commerce brand and whether those agents can complete its checkout, co-founde…
Chen-Ping Yu, whose augmented-reality navigation startup Phiar was acquired by Google in 2022, launched Atmee v2 in early access on November 4th, a real-time avatar model for live video agents that Yu claims renders char…
VLLM introduced a fast-track RFC for managing AI-agent-generated GPU kernels after a Kimi K3 optimization fused decode operations on NVIDIA GB300 GPUs, cutting the kernel count from 23 stock kernels to 13 CuTe DSL kernel…
A developer built Quasnooze, an open-source local AI assistant that audits each model reply against the turn's actual tool trace and memory facts rather than trusting the prose. The guard module, src/guard.js, is a pure …
A developer has proposed a unified AI code governance policy for teams running GitHub Copilot, Cursor and Claude Code simultaneously, arguing that per-agent vendor controls leave policy fragmented across separate rule fi…