AI Research news and analysis on Web Pulse: 8663 curated articles tracking the latest AI Research developments, tools, and research, updated continuously from vetted sources.
Researchers warn that generative AI may weaken critical thinking, memory, and creativity, echoing past fears about the Google effect and calculators. MIT researcher Nataliya Kosmyna argues AI poses a unique cognitive ris…
Geology's June 2026 insurance study found that incumbent insurers capture 78-82% of mentions on ChatGPT, Perplexity, and Gemini, and 66% on Google AI Overviews. The company developed a measurement framework to help insur…
Indonesia's communications ministry on July 7, 2026 emphasized the development of national compute clusters as critical to the country's AI competitiveness, with Deputy Minister Nezar Patria urging stronger domestic infr…
Anthropic announced plans to invest up to $15 billion in Australian data centers by mid-2027, aiming to secure 1.4 gigawatts of capacity for AI model training and deployment. The expansion follows a memorandum of underst…
A new AI tool called Fable – Inner Forward Pass has been released, enabling advanced neural network interpretability through inner forward passes. The tool aims to help researchers understand model decision-making proces…
TechEx Europe 2026 will take place on 19-20 October at RAI Amsterdam, gathering over 8,000 senior enterprise technology leaders across eight co-located events focused on AI, cybersecurity, cloud, and other domains. The c…
Doric released Plotline, a context-integrity benchmark for LLM applications that tests whether systems maintain accurate facts across long, messy sessions. The benchmark exposed failures in models from Anthropic, OpenAI,…
Code maintainability has plummeted in the AI coding era, with code duplication up 81% and code reuse down 70%, according to a GitClear and GitKraken study of 623 million code changes from 2023 to 2026. Legacy refactoring…
In 2026, running large language models locally has become practical and cost-effective for many use cases, with open-weight models matching mid-tier cloud APIs on benchmarks and consumer GPUs capable of hosting 70B-param…
BottleCap released the Thinking CAP model series, a fine-tuned version of Qwen3.6-27B that reduces reasoning tokens by 46% while maintaining benchmark performance. The model aims to cut inference costs and latency by eli…
Sony Interactive Entertainment leadership called AI an "important foundational technology" supporting its strategy, emphasizing that AI tools improve development quality and speed rather than cost efficiency. The company…
Researchers at Aalto University used machine learning to accelerate the discovery of two new kagome superconductors, YRu3B2 and LuRu3B2, with critical temperatures of 0.81 K and 0.95 K. The study, published in Physical R…
A team at the Chinese Academy of Sciences has proposed a Cybersecurity AI Scientist, an autonomous AI agent designed to automate security research, addressing the slow, manual nature of current security research. The con…
Artificial intelligence is forcing society to confront deep social disconnection, as algorithms and platforms intensify fragmentation while offering tools for community rebuilding. AI can revitalize local economies and c…
Microsoft is cutting 3,200 Xbox jobs and spinning off five gaming studios after a $69 billion Activision Blizzard acquisition and $20 billion in expansion failed to generate growth, Xbox CEO Asha Sharma announced. The la…
Nvidia became the first company to reach a $5 trillion market capitalization on October 29, 2025, driven by its early bet on parallel computing through CUDA in 2006. The company's chips, initially used for gaming, became…
The MiMo-V2.5 model family introduces Hybrid Sliding Window Attention (Hybrid SWA) to reduce KVCache storage to roughly 1/7 of Full Attention, sparse MoE activation to cut per-token compute, and multimodal encoders for c…
NVIDIA and Hugging Face are collaborating to integrate NVIDIA's Isaac GR00T 1.7 model and Isaac Teleop framework into Hugging Face's LeRobot open-source robotics library, with NVIDIA Cosmos 3 planned for future release. …
A new study reveals that connecting an AI agent to more than 20 MCP tools causes accuracy to collapse to 13.62%, with excessive token consumption. Researchers found that tool count is not a free upgrade, and reliability …
Tencent released Hy3, an open-source 295B-parameter Mixture-of-Experts model with 21B active parameters and 256K context length, under the Apache License 2.0. The model achieves state-of-the-art results on coding and rea…