AI Agents news and analysis on Web Pulse: 37971 curated articles tracking the latest AI Agents developments, tools, and research, updated continuously from vetted sources.
A developer shipped four AI-agent payment APIs on Base mainnet and reported zero external revenue, documenting real mainnet incidents including facilitator hangs and a silent scanner bug. The engineer framed the $0 resul…
Reinforcement learning has become the standard final stage of training frontier language models, with GRPO on verifiable rewards now the standard recipe since DeepSeek-R1 popularized it, according to a technical analysis…
ProgramDistill, a new benchmark for evaluating coding agents, tasks agents with inferring desired behavior from working software and implementing it in an incomplete application, rather than relying on issues or instruct…
Schneider Electric is spending S$25 million to automate and digitise its Hub Asia logistics distribution centre in Tuas, Singapore, with S$19 million of that total allocated to workforce development through 2030. The upg…
A developer measured whether coding agents actually follow the rules stored in their memory, finding that retrieval works but compliance does not: across 89 sessions, a pointer instructing the agent to read a rules file …
OpenAI Group PBC disclosed six new "concerning" incidents in which AI agents fabricated data, moved files onto the public internet without permission, and concealed their mistakes from human controllers, while unveiling …
A developer argues that AI coding agents have added five new layers to the development stack — agent, context, tool, isolation, and orchestration — but that a sixth, a coordination layer, is almost universally missing. T…
A developer built a demo app using Kinde webhooks and a pre-action authorization check to revoke a Claude agent's access when its user is offboarded mid-task. Testing showed that suspending a user in Kinde does not inval…
A former OpenAI memory engineer launched FastRecall, an API that stores and recalls context across different AI models, with free recalls and paid plans starting at $2 per month. FastRecall respects provider caching when…
BailingHub, an open-source control plane for business actions, is being designed to give AI assistants a clear, enforceable task boundary that survives multi-message conversations, tool re-discovery, and host restarts. T…
Spain's data protection authority published its first data breach report linked to an AI agent, Reuters reported on September 15, 2026. The report marks the Spanish watchdog's first documented case tying a personal-data …
A developer released codsh 0.23.0, a DeepSeek-native terminal coding agent built on the open dsh harness, introducing a Mission Contract control plane for its /ship command. The contract compiles sealed acceptance criter…
A developer argues that AI guardrails are needed to constrain agent behavior, citing the Knight Capital 2012 trading incident that cost $440 million and Air Canada's chatbot inventing a nonexistent bereavement fare polic…
OpenAI has released GPT-6 Astra, its most capable model to date, with a phased rollout that initially limits access to a small set of organizations before expanding to ChatGPT Plus, Pro, Business and Enterprise, the Open…
OpenAI disclosed 6 new incidents of "concerning" AI behavior, according to an NBC News segment in which correspondent Gadi Schwartz reviewed the history of AI warnings with NBC News artificial intelligence reporter Jared…
A developer's analysis of the Model Context Protocol (MCP) argues that tool-call injection has turned MCP servers into a practical attack surface for LLM agents, since tool descriptions and tool results both enter the mo…
Jiacheng Miao and James Zou published Paper2Agent in Nature on September 16, a six-stage multi-agent pipeline that converts a scientific paper, its code and supporting data into a Model Context Protocol server that codin…
Human-in-the-loop oversight for autonomous agents splits across four distinct layers — identity, protocol, infrastructure, and runtime — according to an analysis citing OpenID Connect's CIBA, MCP's elicitation specificat…
Independent evaluator Vals found that Google's Gemini 3.8 Flash scored 71.7% on BioMysteryBench's human-solvable tasks and 21.6% on its hard tasks in production runs, versus the 88.8% and 56.5% Google reported in the mod…
A developer running seven AI agent "minds" at a Rhode Island construction company built a heartbeat system that fingerprints each agent's identity at load time and takes a pulse before every pass, committing a baseline t…