Generative AI news and analysis on Web Pulse: 15590 curated articles tracking the latest Generative AI developments, tools, and research, updated continuously from vetted sources.
WorkContext launched as an open-source, context-aware productivity workspace that connects docs, tasks, and team collaboration, according to its Show HN announcement. The product offers smart @-mentions, action extractio…
A developer launched FormulaAI, an AI-powered Excel tool that parses spreadsheet files locally in the browser using SheetJS and sends only data summaries to the AI, so full files are never uploaded to a server. The tool …
Researchers propose enhancing photogrammetric Digital Surface Models (DSMs) derived from satellite stereo imagery by using pretrained diffusion models with multimodal conditioning, addressing noise, outliers, and voids t…
Imgly published a guide and open-source demo app, editable-ai-ads, that turns a single text prompt into an ad composed of four separate layers — generated background, product cutout, real text headline, and locked logo —…
A developer built NOS Workspace, a tool that renders pasted or uploaded AI-generated HTML as it was authored, allows text to be edited directly on the page without touching code, and shares each document via a link that …
AG Grid shipped AG Studio 3.0, adding a Studio Agent Framework that lets developers connect their own LLM so users can build dashboards and analyze data with natural language while the host application retains control of…
Felix Rieseberg redesigned his personal homepage using roughly 60 parallel Claude threads running entirely in the cloud on Opus 5.5, without writing or running any code locally, opening GitHub, or making a pull request. …
SeedRouter launched a unified API that routes requests to language, image, video and audio models from OpenAI, Anthropic, Google and ByteDance under a single key, with pay-per-request pricing and no monthly minimum. The …
A developer shared three prompting strategies for AI-assisted software development: an incremental story-by-story approach, a single large "one hit" prompt, and a hybrid method that uses a chat session to produce a detai…
OpenVid maintainer Hanish Keloth released an open-source, AGPL-3.0-only tool that lets users build narrated video presentations locally from editable scenes, rendering MP4s on their own machines. The workflow requires Py…
Alibaba's Qwen released Qwen3.8 Omni Flash on 2026-09-17, a multimodal model accepting text, image, audio and video with a 1000k context window and 131k maximum output, according to the provider announcement listed on mo…
A paper titled "FoMo: Forking Moment in Generative Trajectory as a Perceptual Distance" proposes a reference-based image quality assessment (IQA) metric that measures the perceptual distance between a pair of images by e…
A developer who built the AI text-to-video generator CineGen shared lessons from hundreds of test renders on how video prompt engineering differs from image prompting. The core finding is that video prompts must specify …
A Muse AI Agent auto-reply falsely told a buyer "Yep I'm here!" at 9:27, after the buyer, Usman, had waited at the seller's building from about 9:15 and left angry at 9:38, resulting in a negative rating for the MX Keys …
A new arXiv paper (2609.30492v1) reports that evolutionary persona generation increased response diversity by 78.8%, originality by 26.1%, flexibility by 49.5%, and holistic creativity by 13.9% over task-only prompting o…
A new algorithm called Spectral Feedback improves test-time alignment of discrete diffusion models by selecting which token positions to re-mask and re-sample, according to an arXiv paper (arXiv:2609.30456v1). Applied to…
A new inference-time sampler lets pretrained diffusion models construct the central anatomy of the population they synthesize, called the intrinsic atlas, without any retraining or deformable registration, according to a…
Researchers introduced MVAgent, a multi-agent pipeline that recasts multi-shot agentic video generation as condition construction, with agents collaborating through typed conditioning inputs. MVAgent uses a Spatial Groun…
Researchers introduced TrafficImag, described as the first benchmark for counterfactual roadside traffic video generation, combining 9,022 annotated images, 7,043 deduplicated video clips, and 31,145 actor-centered histo…
SlideLab, a training-free multi-agent framework for generating scientific presentations from research papers, was preferred over both open-source and commercial systems on 77% of papers in a blind human preference study …