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.
A developer released Design Decision Method, an MIT-licensed Agent Skill that adds a structured "decision record" step to AI-assisted interface work so coding agents retain the reasoning behind design choices rather than…
A preprint by Adam Manji, "Who Approved This?", documents that at one fintech startup, 71 of 263 non-dependency-bot merges into main during a period of enforced branch protection relied solely on approvals submitted by a…
Axion 1.0, a fully open agent foundation model, was released with a training recipe of only three stages: two lightweight supervised fine-tuning stages followed by reinforcement learning, with no dedicated midtraining st…
A Qwen3-8B-based long-context agent named MEMENTO, trained to manage its own context through compaction, raised book-level question-answering accuracy from 36.9% to 67.8% while never holding more than 20K tokens in conte…
Identical Gemini 3.6 Flash buyer and seller agents reached agreement in 77.3% of Korean-language negotiations versus 52.5% in English — a 24.8-point gap, Holm-adjusted p = .0039 — across 432 attempted negotiations of 12 …
A new arXiv paper (2611.00420v1) introduces Circuit Oracle, an off-the-shelf LLM with a multi-agent harness and task-specific authored skills that answers natural-language questions about a target model by reading its at…
A study on DisasterBench found that majority voting leaves an already-discovered correct workflow unselected on 10-19% of tasks per candidate pool, while structural validation never removed the last correct workflow in t…
A practitioner reports that speech input and output remain underused with AI systems, citing a personal workflow that pairs NVIDIA's Parakeet STT models with the FluidVoice app and a Claude Opus cleanup pass to make loca…
CodingCat.dev wired a 14-repository open-source stack into its production site and a local RTX 5090 video studio to build a 3D Vinyl-Toy AJ the Cat mascot, using Three.js, Blender and Blender MCP (ahujasid/mcp-for-blende…
CompSkill, a skill-retrieval method detailed in arXiv paper 2611.01115v1, raises SkillRet NDCG@10 from 64.0 to 70.9 at one eighth of the key dimension by indexing skills with learned 512-d keys and compiling matched same…
A study of four instruction-tuned small language models (Llama-3.2-1B, Qwen2.5-1.5B, SmolLM2-1.7B and OLMo-2-1B) found that depth pruning collapses function calling and mathematical reasoning in every model, with retenti…
A framework called CuriousTogether, introduced in arXiv paper 2611.01301v1, found that curiosity-guided multi-agent LLM interactions improved performance over non-curious settings on seven of eight tasks spanning reasoni…
An empirical study of 54 self-modification experiments across six language models found that whether and when a model stops rewriting its own system prompt is tied to model identity rather than the optimization mode assi…
A paper posted to arXiv (2611.01415v1) proposes Latent Harness, a method that distills an external agent harness or process-level behavior instruction into a rank-8 activation shift applied at the output of every decoder…
A large-scale trace analysis of 3.7 months of production coding-agent traffic from FreeInference found that cached input tokens account for 97.4% of input-token volume and 72.9% of modeled API cost, even though the media…
Brothers Jason and Deon Rabbass released AXON Developer Edition, a free local AI co-processor written in Rust that lets assistants such as ChatGPT, Claude and Meta AI read and write files, run searches and coordinate tas…
Nirmitee.io published guidance on automating EHR workflows that FHIR APIs cannot complete, comparing API access, RPA bots and computer-use agents. The firm notes Epic's public FHIR R4 capability statement lists 60 resour…
SemiAnalysis released GDPx v1.0, a benchmark of 206 long-running agentic tasks across 11 domains built with 14 partners, and found that the best model, Anthropic's Claude Opus 5.5, passed only 12.0% of tasks at $51.41 pe…
Droneship launched a cloud runtime that runs coding agents such as Anthropic's Claude Code and OpenAI's Codex inside isolated gVisor sandboxes via a single command, `droneship run claude`, with each run receiving its own…
Developer Gregory Bolshakov released peekme, an MIT/Apache-2.0 tool that lets users select text in the output of Claude Code, Codex CLI, Pi, GitHub Copilot CLI or Kiro CLI and press Alt+P (Option+P on Mac) to open an inl…