Full-text search across 17253 articles. Combine with topic and date filters; results sorted by relevance.
A new study systematically decomposing OpenEvolve and TTT-Discover search harnesses into their constituent components and evaluating 30 budget-matched harnesses across 12 model-problem pairs using more than 3.1 million L…
The AI industry is shifting focus from raw novelty to durable interfaces, with the Model Context Protocol locking its largest revision ahead of a July 28 release, AMD unveiling its first 2nm x86 server chip, and GitHub C…
A developer highlights the Kesselman List of Estimative Words, a calibrated set of probability terms for intelligence analysis, as useful for large language models. The list assigns specific percentile ranges to words li…
A new study by researchers (submitted on 27 Feb 2023) demonstrates that locality-sensitive hashing does not guarantee privacy, presenting attacks against Google's FLoC and the MinHash Hierarchy system. The attacks refute…
Researchers have developed a new method for monitoring honey bee colony strength using audio IoT sensors, modulation tensorgrams, and recurrent neural networks that improves accuracy and cross-hive generalizability over …
MIT is spending over $3 million on more than 500 AI surveillance cameras across campus, including academic buildings, residence halls, and outdoor areas along Memorial Drive, with installation from November 2025 to Septe…
Major US teacher unions, including the National Education Association and the American Federation of Teachers, are embedding diversity, equity, and inclusion principles into AI classroom guidance, according to a report f…
Princeton University faculty members Hantao Ji, Egemen Kolemen, Josh Atkinson and Jürgen Hackl have been awarded grants to lead projects in the Genesis Mission, a nationwide initiative announced by the U.S. Department of…
Researchers have derived exact backpropagation from Hamilton's least-action principle, unifying inference and gradient computation within a single variational framework on a doubled phase space. The work, submitted to ar…
AIXI Labs, a new AI safety organization focused on algorithmic information theory and continual reinforcement learning, announced its launch to strengthen the technical case that developing artificial superintelligence p…
Google's AI Overviews and AI Mode are consuming original reporting from publishers and presenting synthesized answers that eliminate the need for users to visit the original sources, according to a New York Times examina…
A controlled study pretraining attention-only decoder transformers (Simple Attention Networks) against standard transformers matched for parameters, compute, and depth found that reallocating the feed-forward budget into…
Researchers have identified a new attack surface against multi-hop retrieval-augmented generation (RAG) agents called Salience Induction, which uses truth-preserving edits to redirect reasoning without injecting false fa…
Nanbeige released Nanbeige4.2-3B, a three-billion-parameter language model using a Looped Transformer architecture that the company says increases model capacity without adding parameters, delivering a capable 3B agent. …
Researchers have developed SEAM-V, a hybrid-decoupled vector execution architecture for RISC-V Vector Extension (RVV) that achieves a geometric-mean speedup of 1.34x across 17 representative kernels compared to an Ara-ba…
Researchers have introduced a set of microarchitecture knobs that dynamically control GPU hardware resources to limit AI performance at runtime, achieving up to 80% performance reduction at 1/8 resource availability. The…
Cisco released Antares, a family of open-weight security small language models (SLMs) designed to pinpoint known vulnerabilities in codebases, with two models—Antares-350M and Antares-1B—now available on Hugging Face. Be…
Serial entrepreneur Qasar Younis announced on July 21, 2026, that his team has completed a major integration effort spanning more than nine years, connecting simulators, data engines, data pipelines, scenario editors, RL…
A new computer science paper recommends that international agreements to limit frontier AI development should include a fixed time period after which a new organization specifies conditions for safe AI development, with …
A model retains oddly specific first-hand accounts during training because they cannot be reconstructed from other data, according to a 2023 study measuring how often models reproduce training text. Retention starts at j…