Natural Language Processing

Abridge, the AI-native healthcare company founded in 2018, is projected to support over 80 million patient-clinician conversations this year across 250 U.S. health systems, saving doctors 10–20 hours per week on clinical…

Together AI has released Violin, an open-source video translation tool that uses speech recognition, large language models, and speech synthesis to convert video content across languages. The tool, powered by the Togethe…

Malte Buettner benchmarked bounding box accuracy for Document AI models using pages from the FlashAttention-3 paper, testing Qwen, Kimi, and Mistral via OpenRouter. The evaluation scored models on coverage, intersection-…

Archivists and historians are turning to large language models to rapidly transcribe centuries-old handwritten documents at scale, a task that previously required specialized training or custom software. Researchers at W…

Granola's AI meeting notes tool enables executive recruiters to capture exact compensation figures and leadership examples from candidate conversations by combining handwritten triggers with automated transcript reconstr…

Granola has introduced a human-in-the-loop AI meeting notes system that lets researchers jot rough notes during conversations while the AI fills in exact quotes and supporting context from the transcript. The approach ai…

Automated meeting summary tools use speech recognition, speaker diarization, and natural language processing to convert audio into structured notes, reliably capturing explicit decisions and action items in low-stakes sy…

Every interaction with a modern large language model is structured as a list of messages, each tagged with a role—system, user, or assistant—that shapes how the model responds and how context is managed across multiple t…

Granola introduced meeting templates and recipes to standardize documentation for repeatable meeting types, allowing users to structure AI-enhanced notes and run saved prompts for extracting insights or generating conten…

Professionals lose hours each week reconstructing specific statements from past meetings, as manual note-taking and generic AI summaries fail to surface exact quotes needed for formal documentation. Granola Chat now allo…

Executive recruiters lose critical candidate details when visible recording bots join calls, as executives become guarded during confidential conversations. Granola, an AI notepad that captures device audio without appea…

Fin, formerly known as Intercom, announced today that it has changed its corporate name to Fin, effective immediately. The company's customer service software platform will continue to operate under the Intercom brand, w…

Patrick Bosek hosted a discussion on the *Behind the Docs* podcast about the future of technical writing, arguing that AI functions best as a collaborator rather than a replacement for human writers. The conversation hig…

Large language models like ChatGPT, Gemini, and Claude operate as next-token prediction machines, taking a sequence of tokens as input and outputting a probability distribution over their vocabulary for what the next tok…

Together AI launched Voice Finder, a search tool that indexes over 600 voices across multiple text-to-speech models to help developers quickly select the right voice for their applications. The tool allows users to searc…

Thinking Machines released Interaction Models, its first AI model, designed to enable fully-duplex real-time conversation where the system can listen and speak simultaneously in micro-turns. The model addresses latency a…

A technical writer at Elastic has developed a tiered system for AI-assisted documentation workflows, using GitHub Agentic Workflows to run quality checks in CI pipelines. The system combines deterministic tools like Vale…

American style uses double typographic quotation marks with punctuation placed inside, while British style uses single typographic marks with punctuation placed by meaning. The typographic apostrophe is recommended for a…

Researchers at BalCapRL introduced a balanced reinforcement learning framework for multimodal large language model image captioning that jointly optimizes utility-aware correctness, reference coverage, and linguistic qua…

A user reduced their RSS feed review queue from 1,920 items to 250 by implementing a two-stage filtering system. The system first removes unwanted content through keyword filtering, then uses Anthropic's API to evaluate …