Natural Language Processing news and analysis on Web Pulse: 2550 curated articles tracking the latest Natural Language Processing developments, tools, and research, updated continuously from vetted sources.
An engineer tested how much of a blog post Anthropic's Claude actually retrieves during AI search, finding that the search stage only returns the title and URL—no body content. The WebFetch stage can retrieve the full pa…
Anfloy built an agentic outbound engine using Apify Actors to capture LinkedIn buying signals, achieving reply rates up to 11% and generating $234,000 in qualified pipeline across three campaigns. The system processes 60…
Amazon Quick and Adobe Marketing Agent have integrated to allow marketing teams to access campaign insights via natural language queries in seconds. The integration uses the Model Context Protocol to connect Adobe's mark…
Posting Machine, a content agency using AI, enables two people to manage 20 LinkedIn accounts for B2B SaaS founders by turning customer calls, meeting notes, Slack threads, and product updates into founder-led posts and …
Joanna Stern tested iOS 27's new Siri AI for a week and found it highly capable, particularly in leveraging personal context from on-device data like Messages and Calendar to provide tailored responses. The AI impressed …
The SEO industry must shift from rank-tracking-style AI prompt measurement to volatility and average response tracking, as traditional methods fail due to AI model updates and personalization. ChatGPT's GPT-5 release in …
Retrieval-augmented generation (RAG) allows local LLMs to access external documents without consuming excessive memory, by retrieving only relevant chunks via a small embedding model and vector database, keeping context …
A developer argues that RAG systems often fail because retrieval returns similar documents that lack the factual evidence needed to support an answer. The post proposes adding an explicit evidence check between retrieval…
Maahir Sharma, a 24-year-old software engineer in Dublin, built an AI agent that called hotels to negotiate rates and perks. The agent handled full conversations and secured better deals, even after a hotel representativ…
Google researchers published a paper detailing a new system, the Scalable Cluster Termination System (S-CTS), designed to detect coordinated generative AI spam by analyzing organizational attack structures and text embed…
On Black Friday 2025, AI-driven traffic to retail websites soared 805%, generating nearly $3 billion in sales, with AI-assisted shoppers converting 38% more often. The winners had clean data and intent-aware infrastructu…
Maahir Sharma, a 24-year-old software engineer in Dublin, built an AI agent using Cursor and Bland AI to call US hotels and negotiate room rates. During one call, a hotel representative suspected the agent was AI, but th…
Speechmark, a new Mac app, provides on-device meeting notes with automatic speaker separation, decisions, and action items, keeping all data locally. The one-time purchase app integrates with Claude Desktop and supports …
A developer outlines a technical playbook for monitoring LLM visibility as AI-powered answer engines reshape content discovery. The approach involves automating API calls to models like ChatGPT and Claude, parsing respon…
Liquid AI released two new retrieval models, LFM2.5-Embedding-350M and LFM2.5-ColBERT-350M, for fast multilingual search across 11 languages. The models, available on Hugging Face, are designed as drop-in replacements fo…
Researchers at AIS2C2 2025 presented a prompt-driven tool-calling framework enabling lightweight open-source LLMs to perform complex multi-step tasks, reducing reliance on large proprietary models and lowering compute re…
The San Francisco Voice Company launched a real-time alert system for video and audio applications, designed for teams handling over 10,000 calls per day. The platform detects acoustic features beyond words, such as frus…
AIClaw, an agent platform designed for tool-using workflows, implements a context compression feature that summarizes long middle sections of agent conversations while preserving critical execution state. The system trim…
A developer built an open-source Claude Code skill called delta-engage that finds potential customers on Reddit and LinkedIn by searching for the pain points buyers voice rather than category keywords. The skill filters …
A developer split a single AI assistant into three specialized agents—Celebi, ProgrammierMinna, and DocMinna—each with a distinct model, personality, and job. The multi-agent system routes tasks via a Telegram interface,…