Generative AI news and analysis on Web Pulse: 11864 curated articles tracking the latest Generative AI developments, tools, and research, updated continuously from vetted sources.
GitHub's monthly commits doubled from 1.4 billion in April to 2.9 billion in August, according to the company's outage postmortem, which attributed the surge to AI agents generating code at machine speed. The 2.9 billion…
A developer argues that people prefer AI-generated art because they already prefer low-quality, formulaic art, suggesting that the rise of AI art reflects existing audience tastes rather than a new aesthetic shift.
A growing number of U.S. public schools are encouraging classroom experimentation with AI chatbots to teach students about their flaws, including hallucination, as part of a new AI literacy push. In Charleston County Sch…
AI agents using Claude 3.5 Sonnet, GPT-4o, and Llama-3.1-70B-instruct rebuilt Future Crew's 1993 demo Second Reality from its original assembly sources into a Rust/WGPU codebase, producing bit-identical 320×200 output an…
DiDi Global Inc. has deployed GenMatch, the first end-to-end generative matching framework for micro-view order-dispatching in ride-hailing, in its production environment. The framework, detailed in a new arXiv paper (26…
Researchers introduced TT-Net, a conditional GAN architecture that replaces per-channel SVD denoising with a two-cut tensor-train decomposition to access cross-channel information, and in controlled comparisons it outper…
A new study from arXiv (2608.19218v1) proposes a time-series retrieval framework that grounds multimodal large language models (MLLMs) in remaining useful life (RUL) estimation, improving prediction accuracy on the C-MAP…
Researchers introduced UniLang, a unified generative framework that extends pretrained large language models (LLMs) to treat machine-native symbols as first-class generative units alongside natural-language tokens. UniLa…
Researchers propose MeanFlow-Transfer (MF-T) and Continuous Adversarial MeanFlow (CAMF) to adapt pretrained diffusion or flow models to new domains with limited data while accelerating sampling. MF-T maps heterogeneous s…
Researchers propose a conditioning mechanism for score-based diffusion models that uses multi-speed joint diffusion of target and condition, enforcing conditioning at inference via a plug-in correction term. The method, …
Researchers propose Stream4D, a method that replaces the static 3D reconstruction critic with a feed-forward 4D reconstruction reward to improve streaming autoregressive diffusion video models, addressing geometric drift…
Researchers propose Block3D, a block-wise diffusion framework for text-to-3D generation that partitions discrete shape-token sequences into contiguous blocks, generates them autoregressively, and jointly denoises tokens …
A new arXiv study (2608.19644v1) finds that analog compute-in-memory (CIM) nonidealities disproportionately attenuate and rotate the classifier-free guidance (CFG) residual in Diffusion Transformers (DiTs), and proposes …
Google's Pixel 11 Pro HiLight LED notification feature is nearly useless out of the box, glowing only for Gemini interactions or calls from favorite contacts in five colors, prompting developer Dhananjay_Tech to create H…
ByteDance's Seedance 2.5 AI video generator now generates 30-second 4K videos from a single prompt with up to 50 reference inputs, enabling one-click filmmaking. The tool supports text-to-video, image-to-video, and synce…
EU AI Act enforcement went live on August 2, 2026, and the European Parliament's March 10, 2026 resolution on copyright and generative AI states that content entirely generated by AI without human creative input should n…
Building AI video agents in 2026 has shifted from single API calls to multi-step systems that plan, generate, check, and fix outputs, with model rankings changing monthly—Runway's Gen-4.5 topped the Artificial Analysis l…
Google's AI Overview feature continues to hallucinate basic facts, as demonstrated by a user's search for Python 3.12 dict union operator syntax, where the overview incorrectly implied that the '|' operator does not retu…
A developer at JPMorgan built an automated content pipeline that generates and uploads videos to nine YouTube channels daily, cutting production time by 72 percent. The system uses local AI models via Ollama/Gemini for s…
Pew Research Center found signs of AI writing or substantial editing on 35% of webpages with publication dates after ChatGPT's November 2022 release, based on a study of 490,000 English-language pages from Common Crawl p…