Generative AI news and analysis on Web Pulse: 15623 curated articles tracking the latest Generative AI developments, tools, and research, updated continuously from vetted sources.
Invideo used OpenAI's GPT-6 Astra to make color grading roughly 3x faster, with a single model handling color correction and grading and generating 50 custom video effects in a day, according to OpenAI. The change folds …
OpenAI published a Harvey case study showing that GPT-6 Astra improves legal document structuring and context awareness, cutting manual drafting time by roughly 40%. The company said the model can produce structured, mat…
Google released a text-to-speech capability under its Gemini 3.8 line, expanding the company's native audio-generation stack alongside existing multimodal endpoints. The release gives developers building voice agents and…
Bitmovin released Bitflix, a fictional streaming network reference implementation built as an MCP App on the Skybridge React framework, letting users browse, get recommendations and watch video by talking to ChatGPT or C…
A developer used OpenRouter to run GLM 5.3 Flash on two Rust functions generated from written specs, paying about $0.03 total for the work. The developer reported the model coded the functions well, worked quickly, and p…
A developer has released the first alpha of ForgeDesigner, a desktop app for macOS (Apple Silicon) and Linux (x86_64) that lets users describe a website in plain language and have an AI assistant generate the pages, Mark…
Representation Autoencoders (RAEs) let diffusion models operate in the feature spaces of pretrained visual encoders, but many off-the-shelf encoders are not optimized for faithful reconstruction and discard fine-grained …
A user of Japanese-English machine translation rated ChatGPT at roughly 80 out of 100 and DeepL at about 60 for one-shot paragraph translation, arguing the technology is very good but not "solved." The user, posting on L…
Anthropic announced on Wednesday that its AI model Claude autonomously discovered a new enzyme system in bacterial DNA that the company said could "represent a new gene editing mechanism" similar to CRISPR, after Claude …
Researchers introduced HYDRO, a target-oriented face de-identification method that combines generative de-identification with a dedicated diffusion process to prevent reconstruction attacks, according to an arXiv paper (…
Researchers introduced Wasserstein-Tilted Flow Maps (WTF), a simulation-free reinforcement learning algorithm for fine-tuning flow-based generative models that achieves higher reward with comparable or higher diversity t…
A dual-module framework called LEGO, combining Legal Expert GraphRAG and expert Chain-of-Thought, reached 40.53% exact-match accuracy on the LawExamQA_Civil benchmark using a Qwen3-8B backbone, according to the arXiv pap…
Researchers proposed LLMAE, a method that repurposes a pretrained decoder-only language model as a continuous text autoencoder by exposing a fixed-length latent bottleneck within its internal activations. Instantiated wi…
Qualcomm unveiled a camera-equipped earbuds concept at its Snapdragon Summit in Maui, Hawaii, on September 24, 2026, with audio brand Cleer showcasing the device, which uses visual sensors rather than cameras for photogr…
A developer recommends having AI coding agents render GUI plans as ASCII art before writing any code, arguing that text descriptions of layouts lead to mismatched expectations. In a real example from Command Center, an a…
The UK Competition and Markets Authority is consulting on criteria that AI assistants such as ChatGPT and Perplexity must meet before they can appear on Google's search choice screens, meaning no AI provider is guarantee…
Daniel Miessler detailed two upgrades to his LifeOS AI stack: Vigil, a module in which every LifeOS subsystem sends periodic updates to a central Vigil database that a separate process reads every 10 minutes and routes t…
Tencent released Hunyuan-A13B, an open-source large language model built on a Mixture-of-Experts architecture that carries 80 billion total parameters while activating only 13 billion during inference, according to the m…
Google has begun rolling out monday.com as part of its Gemini Connected Apps expansion, letting users reference monday.com board data inside Gemini-powered conversations in Gmail, Docs, Sheets and Slides. The September 2…
Google released two text-to-speech models on September 23 — Gemini 3.8 Flash TTS and Gemini 3.8 Flash-Lite TTS — available now as gemini-3.8-flash-tts and gemini-3.8-flash-lite-tts in Google AI Studio and the Gemini API,…