Natural Language Processing news and analysis on Web Pulse: 5534 curated articles tracking the latest Natural Language Processing developments, tools, and research, updated continuously from vetted sources.
Google Cloud announced the general availability of BigQuery Conversational Analytics and the Conversational Analytics API, adding to the existing GA of Conversational Analytics in Looker, with Conversational Analytics in…
Rudrendu Paul and Sourav Nandy released LLMScout, a zero-dependency CLI that runs 21 checks against any website to determine how well AI systems can find and cite it. The tool covers technical SEO fundamentals and GEO si…
Apple plans to release an all-new smart home hub between October and early next year, according to Bloomberg's Mark Gurman. The device will combine a HomePod-like speaker with a 7-inch display, run a new operating system…
DocSlicer, a new structure-aware document parser, achieves 31 pages per second on CPU and scored 0.88 overall on the BizDocBench benchmark, outperforming the next-best tool's 0.70. The deterministic parser handles PDF, D…
Liquid AI released LFM2.5-Encoder-230M and LFM2.5-Encoder-350M, general-purpose bidirectional encoders that match or beat larger models on GLUE, SuperGLUE, and multilingual tasks while running about 3.7× faster than Mode…
OpenAI has expanded GPT-Live in ChatGPT Voice to Edu, Business, and Enterprise plans globally, bringing real-time, full-duplex voice architecture to managed workspaces. The rollout includes two variants—GPT-Live-1 for pa…
Tata Communications and Tata Tele Business Services announced a July 28 collaboration to offer a unified AI platform stack for Indian small and medium businesses, including AI voice agents with dedicated fixed-line phone…
Mount Sinai Medical Center in Miami Beach has activated Epic's ambient documentation tool, Chart with Art, for its inpatient nurses as of July 28, 2026, after physicians began using the same tool earlier in 2026. The too…
A developer building a multi-agent AI system for company LinkedIn pages details the creation of an Orchestrator Agent that manages communication and execution order among independent agents. The agent classifies topics a…
A study by AP found that a guidance sentence instructing models to return an empty tag list for wholly-foreign topics achieved 12/12 accuracy across all three tested models (Opus, Gemini, Sonnet) with no side effects on …
An autonomous agent named ALEMBIC researches, writes, and publishes news articles for a site without human review, running weekly on Mondays via GitHub Actions. The system uses the kimi-k3 model through the Moonshot API …
Collabora Online 26.04, the major annual release for 2026, introduces AI-powered writing assistance in Writer, richer document comparison, multi-page view, and redesigned comment cards, marking ten years of open-source, …
Yelp's AI-powered phone service, Yelp Host, can now handle conversational food ordering by connecting directly to restaurant point-of-sale systems like Toast or Square, the company announced. The expansion includes suppo…
A self-taught AI engineer and amateur linguist claimed a breakthrough in June 2026 on Linear A, the undeciphered writing system of the Bronze Age Minoan civilization, by using AI-built programming scripts to test a hypot…
A team of developers fine-tuned NVIDIA's 8-billion-parameter language model to read legal contracts, achieving a validation Macro-F1 score increase from 0.2956 to 0.8246 and 91.5% exact-match accuracy across six controll…
A new benchmark and baseline study evaluating retrieval-augmented generation (RAG) for French immigration law finds that retrieval improves administrative guidance at both model scales tested. Comparing a parametric LLM …
Researchers propose CADER (Confidence-Aware Dynamic Evidence Reasoning), a training-free framework for adaptive long-video reasoning that uses a logit-margin signal to let high-confidence examples exit early and activate…
Researchers released MedLoCoMo, a medical long-context benchmark for evaluating large language models on patient-specific clinical reasoning over multi-admission dialogues, built from deidentified MIMIC-IV records. The b…
Researchers propose Co-E, a training-free system that synchronizes graph and text working memory for multi-hop question answering, outperforming comparable open-backbone baselines on six benchmarks and rivaling larger or…
A two-stage large language model pipeline using Gemini 2.5 Pro and Gemini 2.5 Flash surfaced 3,460 candidate documentation inconsistencies from 3,000 MIMIC-IV-Note discharge summaries, affecting 69.7% of admissions, acco…