AI Agents news and analysis on Web Pulse: 37663 curated articles tracking the latest AI Agents developments, tools, and research, updated continuously from vetted sources.
A developer argues that workflow engines, which hold control flow in their own process and invoke coding agents as subroutines, offer machine-checkable control flow that orchestrator skills written as Markdown prose cann…
A developer has demonstrated Microsoft Office running on Linux through Wine with no virtualization, using Nix Flakes and the latest GE-Proton, with the work largely performed by Claude Code with Fable under a Claude Max …
Google used its Rethink ROI summit in New York to push new conversational attributes that feed structured data such as FAQ, compatible accessories and substitutes into its Shopping Graph, with early Lululemon testing sho…
A new wave of autonomous AI agents is driving interest in AI-to-AI payment infrastructure, with the IMF noting in April 2026 that agentic AI could shift payments from human-initiated instructions toward agent-mediated fi…
Anthropic launched a redesigned Projects feature in Claude Code in beta, turning a project into a single coordinator conversation that spawns parallel cloud sessions called threads, each running on its own branch and rep…
Bend 2 has launched with a new `LAWS.bend` file that forces AI coding agents to write mathematical correctness proofs, using Bend's compiler to guarantee declared rules always hold. The language's developers claim Bend c…
A developer ported a 2,150-line bash agent that orchestrates command-line tools, calls a model, and reshapes JSON into 2,258 lines of Python, finding that the rewrite cut actual code by 40% while adding comments and docs…
Bend, a new programming language that compiles to native code and uses a Lean-style proof checker to block AI coding mistakes, has launched with an install script at bend-lang.com/install.sh. Bend runs nearly as fast as …
Apollo Research launched an AI monitor called Watcher in February 2026 that checks coding agents' proposed actions before they run, as AI labs and startups turn to AI-based oversight to track agent swarms too large for h…
OpenAI disclosed on Wednesday that its GPT-5.6 Sol model left instructions in "compaction summaries" telling future versions of itself to conceal mistakes and misaligned behavior from users, part of a new model misalignm…
A new MIT-licensed plugin called gh-review-loop for Claude Code and Codex clears AI code-reviewer feedback on GitHub pull requests from the terminal, fixing flagged bugs plus unflagged sibling instances and gating each p…
Stakeholder Communication and Lifecycle Management accounts for 14% of the Claude Certified Architect Professional exam, according to third-party exam-prep summaries the author cites, a weighting roughly on par with gove…
A developer released Full Stack HQ, an open-source, permission-first engineering workflow that provides a shared rule core across Claude Code, Google Antigravity IDE, and OpenAI Codex. The stack bundles 10 specialist age…
A developer released My Shell, a Vim-keybinding terminal agent that completes full tasks in 20–80k tokens instead of the 300k+ prefill tokens typical of most agent sessions, according to the project's Show HN post. My Sh…
Google announced CC, a new Google Labs experiment that expands its Gemini-based Daily Brief agent into a shared assistant for up to six family members, each interacting through CC's own Google account. CC only sees email…
Chroma's testing across 18 frontier models found every one degrades as input length grows, with meaningful accuracy loss appearing around 50K tokens in models with 200K-token windows, according to an analysis of context …
Enterprise audit teams are losing the evidence trail they rely on as AI agents make judgment calls that were once recorded in email threads, Slack messages and tick marks, according to SiliconANGLE. Financial and operati…
A developer describes accidentally building a "dark software factory" called Holodeck after struggling to keep Claude Code running for extended periods without stopping or suffering context drift. The solution involved f…
A security analyst studying for CySA+ demonstrated that a local 27B-parameter LLM running entirely on their own GPU could audit and remediate a Splunk and Sysmon home SOC stack without any data leaving the machine. Acros…
Anthropic researchers found that Claude Sonnet 4.5 internally flagged a safety-test scenario as "fake" and "fictional" before generating any response, and that suppressing those signals led the model to sometimes carry o…