Updated daily708 verified stories
Research News tracker.
Material AI papers, studies, datasets, evaluations, and scientific advances with direct links to the underlying research.
- 30stories this week
- 708sources cited
- 708verified timeline entries
Verified sourcing
Every story links to the strongest available evidence: original sources when available, otherwise clearly attributed reporting.
Plain English
What happened, why it matters, and what to watch — without the jargon.
No filler
When the signal is thin, we publish nothing rather than padding the feed.
More stories
9 storiesInnovation
Paper Says Agent-Aware Cache Management Cuts First-Token Delay Up to 45% in Multi-Agent Serving
A new arXiv preprint describes CacheScout, a layer built on the open-source vLLM server that decides what to keep in a model's key-value cache based on which agent is likely to run next. The authors report double-digit latency and throughput gains; the workloads, models, and hardware are not stated in the abstract.arxiv.orgInnovation
Audit of an OpenAI AI-Generated Proof Finds a Reversed Condition, and Publishes a Repair
Two researchers say a lemma proof in Chapter 6 of OpenAI's mathematics document has a polarity error: a test in terms of average success where the next step needs a large conditional failure. They give a counterexample and a corrected proof, and caution that this is not verification of the chapter's main theorem.arxiv.orgInnovation
Replication Study Says FLOPs Still Mispredict AI Runtime, and the Proposed Fix Fails on Newer Hardware
A preprint by two researchers reproduces an earlier study on why equal FLOP counts do not mean equal execution time. It confirms the underlying claim but reports that the α-FLOPs correction formula generally underestimates runtime on newer hardware, which shows jumps and oscillations the formula does not capture.arxiv.orgEnterprise
Benchmark Paper Finds Four Ways to Query Enterprise Data With LLMs All Score Under 26%
A new arXiv preprint pits four architectures for natural-language querying of enterprise databases against each other on a synthetic bilingual benchmark. None answered more than about a quarter of cases correctly, and the design that scored highest was not the safest or the cheapest.arxiv.orgInnovation
Paper Proposes Grading AI Security Agents Without Labels by Measuring Convergence to a Stronger Model
A new arXiv preprint argues security teams can judge whether a memory- or retrieval-equipped AI agent is learning by measuring how far it closes the gap to a stronger "teacher" model, rather than on labeled benchmarks that are often scarce or stale. Judging by a similarly powered model gave no usable signal.arxiv.orgInnovation
New Benchmark Tests Whether AI Assistants Can Remember a Year of Phone Use
A 17-author technical report posted to arXiv introduces MobileMem, a benchmark and framework for on-device long-term memory built from a year-scale collection of mobile experiences. The abstract describes the design but reports no scores, and key details about the underlying data remain undisclosed.arxiv.orgInnovation
Paper Reports Brain-Like Modular Organization Emerging Inside Large Language Models
A new arXiv preprint says large language models develop functionally specialized internal structure that lines up with distinct human brain networks, based on circuit analyses across 46 tasks in four cognitive domains. The abstract page leaves key methodological details unstated.arxiv.orgInnovation
Paper Finds Late Layers of a Mixture-of-Experts Model Tolerate Heavy Expert Masking
A preprint reports that disabling low-magnitude experts in the last five layers of a 35-billion-parameter Mixture-of-Experts model preserved far more usable code-translation outputs than spreading the same cuts across all layers. It covers one model and one benchmark, and the abstract reports no unmasked baseline.arxiv.orgInnovation
New Benchmark Finds AI Agents Wrongly Block Approved Work 28% of the Time
A preprint introduces SteerBench-Work, a 106-scenario test of the moment an AI agent decides to act or pause for review. Across 30 model conditions, the authors report that wrongly holding cleared work was roughly 28 times more common than wrongly allowing unsafe work.arxiv.org
One useful briefing each week
Keep up with AI without living in the feed.
Get the week’s verified AI news, original data, useful tools, learning picks, and fresh AI jobs.
Reach people who are learning AI
Hiring an AI professional or launching a useful AI product? Put it in front of people who came here to learn and act.
Post an AI jobSubmit an AI tool