Natural Language Processing news and analysis on Web Pulse: 5440 curated articles tracking the latest Natural Language Processing developments, tools, and research, updated continuously from vetted sources.
EXL launched a Medical intelligent document processing (IDP) solution built on AWS that combines its Xtrakto.AI document-extraction application with the EXL Insurance LLM to cut the more than 100 minutes per case insuran…
OpenAI's GPT-Live-1, a full-duplex voice model that can delegate reasoning and tool use to a backend agent, requires a typed "delegation contract" to keep voice agents from talking over customers, repeating actions, or l…
TypeSafe AI's Jev is an AI model built for fast, structured decisions rather than open-ended text generation, which the company calls a "System One Model" and distinguishes from conventional large language models. The mo…
Yandex released AliceAI-Foundation-80B-A3B-Base, an 80 billion-parameter large language model, on Hugging Face under an Apache 2.0 license for commercial and academic use. The Mixture-of-Experts model activates 3 billion…
A developer built EduVetta, a SaaS tool that uses AI to automate clerical teaching tasks such as generating lecture summaries, drafting question papers with answer keys, collecting assignments, and creating quiz games. T…
Researchers introduced Gricea, an open-science platform that represents conversational AI (CAI) studies as configurations to address fragmented reporting that hinders replication and knowledge accumulation. The platform …
A developer ran four text-analysis trials on the AIDBDeveloper platform to measure how small workflow design choices affect token usage, caching, and cost across several thousand API requests and millions of tokens. The …
Seeitflow published an animated visual explainer on how tokenization works in large language models at seeitflow.com/ai/ai-foundations/tokens-explained. The post drew 1 point and 0 comments on Hacker News.
Infineon is promoting a multi-sensing approach to intelligent human-machine interface (HMI) design in which edge AI — on-device speech recognition and small language models — replaces cloud dependence and touchscreen-onl…
An open-source Python library called ChunkRank is being submitted to COLING 2027, and its developer is seeking an arXiv cs.CL endorsement and an ARR service contributor from the Hugging Face community. ChunkRank provides…
Legend.org, a New York-based AI company that purchased the nonprofit Mastery Transcript Consortium in July, is developing AI tools to give teachers and students immediate feedback on papers, exams, recorded presentations…
A developer integrated Octic, an on-device AI recorder, with the open-source Octo collaboration platform to automatically turn spoken meeting action items into assigned tasks. The five-stage pipeline runs on-device ASR, …
LanguageOps has built automated checks into its translation editor to catch mechanical formatting errors such as misformatted placeholders like `$0` and `{name}` before reviewers see them, according to the company's guid…
InfoWorld published a list of 20 prompt styles and structures for getting better results from large language models, including step-by-step instructions, few-shot examples, Q&A-style questioning, and prompts that ask the…
Developer lukstei released slop-grader, a CLI tool that grades text files against custom rulesets using the Jev System One model from typesafe.ai. The tool batches lines into groups of 255 and evaluates each rule across …
Yandex released AliceAI-Foundation-80B-A3B-Base, an 80B-parameter mixture-of-experts language model trained entirely from scratch on roughly 18T tokens and published on Hugging Face under Apache 2.0. The base model activ…
A new non-profit site, pressaudit.org, is using Alibaba's Qwen 3-32b large language model to process thousands of news articles and analyze them by region, public interest, and partisan lean, according to the project's l…
A developer who has reviewed recordings from live AI voice agent deployments argues that callers hang up not because of voice quality or speed but because agents are scoped to intents they cannot finish and cannot hand o…
A developer built a ~10.6 million parameter decoder-only Transformer from scratch in PyTorch and trained it end-to-end on the Tiny Shakespeare dataset using only the free tier of Google Colab's T4 GPU. The 6-layer, 6-hea…
A developer has introduced YOalphabet and YOconlang, a spatial 4-bit bytecode script and constructed language that maps visual glyph geometry directly to binary hardware registers, eliminating the need for vector embeddi…