Generative AI news and analysis on Web Pulse: 15500 curated articles tracking the latest Generative AI developments, tools, and research, updated continuously from vetted sources.
KittenTTS 2, released by KittenML, is a 1.7-billion-parameter speech language model that clones a voice from as little as 5 seconds of audio, a jump from the original KittenTTS models' 15M to 80M parameters. The new vers…
An agile coach built a production-ready team assessment application in about an hour using Replit's AI vibe-coding platform after an initial attempt failed, succeeding only when he described the desired outcome and user …
A solo developer built SideDog, an AI browser side-panel assistant that summarizes web pages, PDFs and videos, translates pages and answers questions about screenshots, in about a month while pairing with an AI coding as…
A developer published a practical guide for ML engineers on building AI-generated image detection pipelines, organizing detection techniques into four families: signal-level methods, metadata and provenance checks, learn…
A developer working on the Nano Banana 2.1 image-generation adapter argues that automatic retries on a timed-out image-generation request can trigger a second paid job, since a provider may have already accepted the orig…
Mistral AI released Mistral Large 4, its newest flagship model, as an API public preview through Mistral Studio, with open weights planned for the end of October. The model uses a granular mixture-of-experts design, acce…
Yandex, Russia's largest internet technology company and operator of the YandexGPT generative AI model, said a suspected Ukrainian drone attack this week caused a fire at one of its major data centres and forced it to fu…
Unity announced Unity Spark, a generative AI tool that lets developers prompt a game into existence using assets from the Unity Asset Store, following the open beta of Unity AI in May 2026. Unity President and CEO Matt B…
Cynxio's Derwin Oesmanto directed "Carry Me," a music video animating human history as a 384x384 swarm of 147,456 tiles, with animation code written in TypeScript using Claude Opus 5.5 in Claude Code and motion rendered …
Developer franzenzenhofer released whitedeck, an open-source Claude Code skill and CLI that converts markdown into Apple Keynote-style white-theme slides in HTML, PDF, editable PPTX and native .key formats. Whitedeck ext…
A browser-based first-person shooter called Claude OF DUTY was built with no art assets, generating all nineteen surface textures, soldier meshes and sound effects procedurally at runtime. The game implements a hand-buil…
Flathub replaced its near-total ban on generative AI app and game submissions with a disclosure policy, initially amended on September 4th, requiring submitters to identify any AI-generated code, documentation, packaging…
OpenAI's newly unveiled GPT-6 introduces an "Intelligent UI" that replaces the traditional chat box with a spatial, interactive environment, including a virtual fitting-room feature that renders clothing on a personalize…
Daniel, developer of the BF-JIT brainfuck just-in-time compiler, reported that Claude Opus 5.5 optimized the 23MB brainfuck ray tracer ray.bf to roughly a 9000x speedup in 1 hour and 45 minutes, cutting a full render fro…
AI crawlers accounted for 23% of the 54,367 counted bot hits on mojodojo.io over the last 30 days, with 12,666 hits split across AI training (8,841, 16.3%), AI search (2,829, 5.2%) and AI user fetch (996, 1.8%), accordin…
A developer's cost and reliability breakdown of long-form text-to-speech finds that rendering 50,000 words runs roughly $4.50 on OpenAI TTS-1, $4.80 on Azure's direct console, and $66–$85 on ElevenLabs' Creator tier once…
A developer added browser-local depth estimation to Timeline Studio, an open-source video editor, using Depth Anything V2 Small run through Transformers.js with WebGPU and a Q4F16 configuration in a Web Worker. The imple…
OpenAI shipped GPT-6 with Intelligent UI on October 7, adding the ability to generate interactive calculators, sortable tables, and planners inline in ChatGPT. The release shifts ChatGPT output from static text toward ge…
Anthropic released Claude Haiku 5.5, a small model priced roughly 90% below Haiku 4.5 for prompts under 100K tokens and matching GPT-6 Luna's price at that tier, with adjustable effort levels (Low through Max) added to t…
A developer built FoliageStride AI, an open-source browser-based running route generator that uses Google's Gemma 2 (gemma2:2b) model through Ollama to synthesize scenic autumn routes from distance, terrain and pacing pr…