News

A study evaluating six large language models within the digital mental health intervention Wysa found that retrieval-augmented generation (RAG) improved intent detection accuracy, recall, and precision compared to RAG-di…

A forensic reproducibility audit of a chest-radiograph vision-language model benchmark found that 297 of 300 planned model-prompt calls yielded nonempty reports, but 60 Claude calls labeled A/B were executed with the sam…

A new benchmark called HANDBOOK.md shows that leading language-model agents fail to follow long, binding policy documents in enterprise settings, with the best model passing only 36.2% of trials under strict grading. The…

A new study from arXiv introduces a controlled benchmark revealing that the site of adaptation in transformers—early, middle, or late layers—produces distinct learning signatures across five objectives: lexical binding, …

A new open-source benchmarking framework, Inspect India Evals, reveals that large language models (LLMs) tested on Indian linguistic and cultural contexts show significant performance gaps, with Sarvam-M 24B and Gemma 2 …

Researchers released a human-in-the-loop corpus for LLM-based simplification of scientific summaries, finding that GPT-4o-mini-generated summaries were preferred for comprehensibility and simplicity but that expert edits…

A new theoretical framework from a paper on arXiv (2607.25507v1) proposes a bounded spectral analysis of rotary phase alignment in Transformer language models, proving that uniformly bounded phase displacement limits deg…

Researchers present VLD-RAG, an agentic multimodal retrieval-augmented generation framework for question answering over visually-rich long documents, achieving improved evidence-page retrieval and end-task performance on…

A new study from arXiv preprint 2607.24797v1 finds that decoder-only large language models (LLMs) such as GPT-2, OPT, and Pythia entangle reading and writing into a single code, unlike the human brain's double dissociati…

Cardiologent, a multi-agent clinical decision support system, outperforms cardiologists and large language models in patient-level arrhythmia assessment, urgency, and management by integrating ECG and photoplethysmogram …

A new uncertainty-aware framework, OpenCoder, improves GPT selected-output correctness in repository-level code generation from 56.25% to 78.13% on the RepoExec-inline benchmark, according to a preprint on arXiv (2607.24…

Snyk launched the Snyk Connect Fan Zone tour across 8 cities and 3 virtual sessions, combining AI security workshops, networking, and game-day rallies to build practical AI security skills. The tour featured hands-on cha…

A new analysis suggests that large language models (LLMs) may be able to find security flaws in formally verified software like the seL4 microkernel by targeting gaps between mathematical proofs and physical hardware exe…

A Cell study published July 27 introduced Oncoformer, a multimodal transformer trained on routine health records, lab tests and chest X-rays from 3.67 million people, achieving an AUROC of 0.869 for predicting cancer up …

OCBC will use an agentic AI platform called Helios to cut wealth client onboarding time to 15 business days, down from the current industry median of about six weeks. The platform, built by the bank's compliance team, mo…

Researchers propose Transformer Transformer, a unified diffusion model that generates complete robot embodiments—every link, joint, motor, and inertial property—optimized for a given manipulation demonstration. In a clot…

Anshul Raturi, a second-year computer engineering student at Pithuwala Polytechnic, built and launched two AI platforms: RaturiHub AI and MAX AI Assistant. RaturiHub AI is now live on ProductHunt.

A Hacker News post speculates that if U.S. labs slow down AGI development, a scenario could unfold by 2028 where Chinese AI labs such as Moonshot Kimi 6, Alibaba Qwen 5, Z.ai, and MiniMax claim AGI-level capabilities, wh…

A thought experiment proposes a dystopian scenario where large language models cannot write code but can generate software binaries, drawing parallels to how non-software workers experience AI advances. The piece referen…

Mozilla's privacy analysis of dating platforms, though a few years old, remains educational. Happn may require facial verification, while Eharmony uses AI to scan every message, raising privacy concerns for users.