AI Research news and analysis on Web Pulse: 22666 curated articles tracking the latest AI Research developments, tools, and research, updated continuously from vetted sources.
Moonlake founders Chris Manning and Fan-yun Sun are proposing causal, action-conditioned world models that fuse game engine abstractions, symbolic reasoning, and learned neural priors, positioning the approach against Ya…
Researchers at the Indian Institute of Technology (IIT) Bombay demonstrated an ultra-low-power neuromorphic neuron fabricated in standard GlobalFoundries 45-nm CMOS technology that consumed about 1,000× less power and sh…
An experiment pitting five diverse AI agents against five identical clones of the same model (Opus 5) over 14 nights produced a pooled Brier score of 0.0225 for the diverse agents versus 0.0275 for the clones, with the d…
DeepSeek released DeepSeek V4.1 Flash on Thursday, a 763 billion-parameter model that is more than 2.5x the size of the model it replaces and larger than the V3 and R1 models that brought the Chinese AI company attention…
Researchers introduced TempCloze, a video cloze benchmark designed to evaluate visual temporal reasoning in Video-LLMs while reducing linguistic shortcuts from option wording, answer correlations, or language priors. The…
Late-interaction retrieval remains the state-of-the-art for visual document search but its accuracy comes at a storage cost, since existing compression methods retain only a subset or local average of the roughly 1,000 v…
Researchers introduced X-AuT, a progressive framework that selects layer combinations to compress audio-encoder depth for speech large language models, addressing the deletion and premature end-of-sequence errors caused …
A developer argues that most AI agent projects should instead be built as conventional software that calls language models at a handful of predefined points, reserving true agents only for tasks whose steps cannot be enu…
Google Research released ToolGrad, an ACL 2026 Findings framework that generates tool-use datasets by building a verified API chain first and then writing the matching user query, reaching a 99.8% pass rate on ToolBench …
Anthropic accused Alibaba, Moonshot, and DeepSeek of running large-scale distillation campaigns against Claude, alleging roughly 200 million exchanges were used to extract the model's reasoning traces for training compet…
A researcher announced they are joining the OpenAI nonprofit board and its Safety and Security Committee, warning that rapid AI capability gains could lead to a catastrophic and irreversible loss of control in the near t…
Red Hat senior machine learning engineer Bartowski published a per-tensor layout map method for GGUF quantization on September 10th, replacing fixed name-, shape- and position-based precision rules with maps derived from…
Liam Fedus, OpenAI's former VP of Post-Training, unveiled Periodic Labs' blueprint for an "AI Foundation Lab for Atoms" in a conversation on the No Priors podcast, arguing that scaling laws must expand from digital code …
A developer reports that Qwen 3.8 27b, a locally run open-source model, now matches frontier cloud models like Claude Opus 4.6 for agentic coding tasks, scoring 34 versus 32 on the Artificial Analysis index at the xhigh …
McKinsey's State of AI 2026 survey found that 32% of organizations skipped a software purchase this year because an AI coding agent could build the equivalent in-house, according to the report's coverage of 1,719 organiz…
OpenAI filed a patent for a second AI model designed to read another model's chain-of-thought reasoning and flag reward hacking, misgeneralization, sycophancy, and deception, according to a weekly roundup of 272 Big Tech…
Researchers Ayhan Suleymanzade and co-authors submitted a paper to arXiv on 10 September 2026 proposing "looped flows," a training method that uses local denoising objectives with progressively decreasing noise levels an…
A field study published September 10, 2026 found that asking an AI agent to commit to moral integrity cut forbidden reads of an out-of-bounds solution document from 33 of 200 runs (16.5%) under a naive prompt to 10 of 19…
OpenAI's GPT-6 Astra became the first AI model to score 100% on ExploitBench, a Carnegie Mellon University benchmark that grades how far a model can carry an exploit chain to arbitrary code execution, earning a "Critical…
OpenAI's Astra reached a nether fortress in Vals AI's long-horizon Minecraft computer-use evaluation, a milestone Vals AI said no AI has achieved in history. Vals AI stated Astra is now in the final stretches of beating …