AI Research news and analysis on Web Pulse: 22662 curated articles tracking the latest AI Research developments, tools, and research, updated continuously from vetted sources.
Elon Musk said on September 11 that xAI has delayed Grok 4.7, its next-generation AI model, for "a few more days" of reinforcement-learning tuning, pushing back a release he had previously pegged four separate times sinc…
Researchers from HHMI Janelia, Google Research, and other collaborators used AI to produce a complete brain map of a male fruit fly, a years-long project announced in a Google Research blog post. Software engineers then …
Julia Stadlmann of the University of Illinois Urbana–Champaign set a new record on the twin prime conjecture on the last day of August, the first advance on the problem in more than a decade, only to be surpassed within …
OpenAI has claimed a solution to alternatives C and D of the Navier–Stokes Millennium Prize Problem, according to a blog post analyzing the result. In the same week, two other groups announced finite-time blowup results …
Shiladitya Banerjee, a computational biophysicist at Georgia Tech's School of Physics, ran an experiment with thousands of individual E. coli cells in which the bacteria behaved differently depending on the previous food…
A Hacker News user asked how successive generations of AI models are built, questioning whether Anthropic's Opus is trained from scratch, derived from Sonnet, or largely the same model with more compute and training. The…
OpenAI deployed roughly 10,000 AI agents simultaneously in August 2026 and solved the Navier-Stokes Millennium Prize Problem in 88 hours, according to The Verge. The result highlights a broader shift in AI agent architec…
Google will establish a physical space on Ohio State University's west campus within its Innovation District, deepening an AI-focused partnership the university calls "one of the largest foundational partnerships between…
A white paper titled "Visual General Intelligence: A White Paper," published on arXiv as 2608.25924 by more than 21 researchers from Google DeepMind, Harvard, and other institutions, argues that images, video, and geomet…
Oriol Vinyals, until recently head of research at Google DeepMind, said AI self-improvement will accelerate research by a factor of ten but will not trigger a sudden intelligence explosion, citing two bottlenecks: genera…
Researchers introduced ReasAlign, a model-level defense against indirect prompt injection attacks that uses structured reasoning steps and a test-time scaling mechanism with a preference-optimized judge model to score an…
Researchers presented ReasoningBomb, a reinforcement-learning-based prompt-induced inference-time denial-of-service (PI-DoS) framework that drives victim large reasoning models into pathologically long reasoning traces u…
OpenDiscoveryTrace released a dataset of 558 full AI-agent trajectories across 124 scientific tasks in drug discovery, genomics, and materials science, hosted on arXiv as 2609.09203v1, to expose how output-only benchmark…
A study by Off-by-1 Labs, the security team launched by 1Password, found that frontier AI models produced complete, behavior-preserving security patches only about 26% of the time across 6,080 graded attempts. Testing Ch…
A position paper proposes system-level defenses against indirect prompt injection attacks in AI agents powered by large language models, arguing that dynamic replanning and security policy updates are often necessary for…
Ketney Otto, a researcher and academic staff member at Lucian Blaga University of Sibiu in Romania, is requesting a cs.LG arXiv endorsement for a manuscript titled "Leakage-Aware Benchmarking of Machine Learning for Food…
Researchers proposed Onyx, a disk-oblivious approximate nearest neighbor (ANN) search system that achieves 1.7–9.9× lower cost and 2.3–12.3× lower latency than the state-of-the-art oblivious ANN search system. Onyx inver…
OpenAI's GPT-6 Astra became the first AI model to surpass human-level performance on all five Drone-Bench subtasks — 3D mapping, self-localization, route planning, person detection, and motion tracking — in at least one …
Humans& released Persimmon, a large-scale "user model" designed to simulate how people behave in multi-turn, multi-user chat settings, making the model available through a playground and API as a research preview. The co…
A post on the Alignment Forum reports that Astra, an AI system, can perform a concerning amount of work without producing any chain of thought. The post, titled "Astra does a concerning amount of work with no chain of th…