AI Agents news and analysis on Web Pulse: 49136 curated articles tracking the latest AI Agents developments, tools, and research, updated continuously from vetted sources.
A developer rebuilt a Grok-generated Kalshi prediction-market paper-trading bot on Nov 5, 2026 and found its HTTP edge mishandled rate limiting: on a 429 with Retry-After: 3, the bot retried after roughly 0.5 seconds of …
A developer launched Botel, a pay-as-you-go hosting service for AI agent harnesses such as openclaw that runs each agent in a Firecracker-based Linux VM and charges only while the agent is awake, at 3 cents per GB of RAM…
A paper submitted to arXiv on 5 Nov 2026 presents Cognitive Tours, a web-based prototype for creating and following curated paths that connect user needs, design decisions, code, tests and open questions, aimed at mitiga…
Anthropic confirmed that Claude Code sometimes placed the signed-in user's email address in the User-Agent header of web requests the model wrote itself, such as curl calls to sites requesting a script contact. Version 2…
A November 5, 2026 arXiv paper introduces SwarmEvolve, a framework in which hundreds of low-cost LLM agents per round reason in latent space without decoding to text and share information by passing KV caches. On 21 prob…
Anthropic contributor mpoteat confirmed that Claude Code 2.1.317 and later ship a built-in mod called cc-plugin-privacy, on by default, after GitHub issue #107757 reported on November 5, 2026 that Claude Code 2.1.318 on …
A November 5, 2026 arXiv paper introduced XStack, a benchmark that grades AI agents on migrating machine learning repositories from PyTorch on NVIDIA GPUs to JAX on Google TPUs, with each agent given a PyTorch repository…
Benjamin Mayanja built Soma, a self-hostable learning and publishing platform that combines a private markdown notebook, an editorial reading experience, and a one-toggle public publish option, with all writing stored in…
A developer documented a reusable Search → Extract → optional Render web-research workflow for Claude Code built on the AgentSearch MCP server, which exposes five tools including agentsearch_web_search, agentsearch_extra…
Walmart said on October 21 that it is building the Walmart Agent Gateway, an invite-only beta running on Anthropic's Model Context Protocol that lets AI agents shop and check out on Walmart's platform, with full rollout …
Hugging Face released a set of "harness adapters" — one LoRA adapter per model and agent harness, trained with reinforcement learning inside unmodified pi, opencode and hermes harnesses — with every tool call inspected b…
CoinGecko cut the price of its experimental x402 pay-per-request API endpoints from $0.01 to $0.005 USDC per call, a 50% reduction, and added USDC settlement on Polygon alongside existing support for Base and Solana. The…
In episode two of The Friday Deploy podcast, hosts Garrett and the author described a workflow for keeping AI-assisted pull requests reviewable by bounding review loops: after the first review, review only the fixes rath…
Developer Krish Punia built OpsSwitch, an open-source multi-agent incident-response system on the Switch AI platform that runs a four-agent squad inside team chat to triage, investigate, patch, and document production in…
A developer built ReproRoom, a hackathon workflow that pairs a Mattermost chat room with a local Codex coding agent to reproduce, fix, and verify a deliberately planted checkout bug in a Node.js app. Given a report speci…
Equibles, an MCP server that supplies ChatGPT and Claude with SEC filings and market data, detailed three data-handling traps that can mislead AI assistants analyzing Form 4 insider trades. The walkthrough, using GameSto…
Airbyte released an MCP Server for Airbyte Cloud and Airbyte Enterprise Flex that lets users query data in natural language from Claude or ChatGPT and use AI agents to monitor, diagnose and fix data pipelines, the connec…
GPT-6 Astra High won a 10-model round-robin Pokémon Showdown tournament, taking 75 of 90 games, while Qwen 3.8 27B finished last, according to an experiment run by the cofounder of an inference company who spent $1,000 o…
Constructive, the company led by Dan Lynch, released Safegres on November 6th, an open-source read-only verifier that grades Postgres grants and row-level security in CI. Safegres is aimed at teams whose AI builders and …
A developer released Current, an open-source Elixir tool that gives AI coding agents change control by restricting them to producing proposals rather than acting directly, with nothing applied until a human approves. Cur…