News

50 items (filtered)

The raw AI news stream — everything we ingest across RSS, Hacker News, arXiv, GitHub, Bluesky, YouTube and more, newest first.

24h 48h 7d 30d 90d All (6mo)

Source	Age	Title	Summary	Status
bluesky bluesky-ai	1mo	*In my early post-marital life, I assumed there would be gigolos, young men eager to spend my money. I now realize that this old house that…	*In my early post-marital life, I assumed there would be gigolos, young men eager to spend my money. I now realize that this old house…	↗Not briefed
arxiv arxiv-ai	1mo	DGF-Bench: A Benchmark for Simulating and Auditing Deception Against Multi-Agent Governance Boards	Tool-using language-model agents can review enterprise projects as governance boards do: they read the evidence, apply written rules a…	↗Not briefed
hn hn-ai	1mo	Lean formalization of the Hamilton-Perelman proof of the Poincaré conjecture	—	↗Not briefed
bluesky bluesky-ai	1mo	Paper on how to vibe-code replication studies in economics www.nber.org/papers/w35782	Paper on how to vibe-code replication studies in economics www.nber.org/papers/w35782	↗Not briefed
hn hn-ai	1mo	Understanding LLMs with an Interactive Tutorial	—	↗Not briefed
rss ai-snake-oil	1mo	AI existential risk probabilities are (still) too unreliable to inform policy	How speculation gets laundered through pseudo-quantification	↗Not briefed
arxiv arxiv-ai	1mo	Don't Throw Away the Tail: Action Upcycling for Policy Acceleration	Modern robot policies predict a chunk of future actions from a single observation, execute only a prefix, and discard the rest before…	↗Not briefed
bluesky bluesky-ai	1mo	1. These companies steal research. I don’t think there’s any other way to n out this. OpenAI and Anthropic use scientists’ own work against…	1. These companies steal research. I don’t think there’s any other way to n out this. OpenAI and Anthropic use scientists’ own work ag…	↗Not briefed
hn hn-ai	1mo	Nodd – Tiny, use-case-specific decision models that can run in browser	—	↗Not briefed
bluesky bluesky-ai	1mo	Von einem verdammten LLM. In 99% der Fälle wenn über AI/KI gesprochen wird geht es tatsächlich um LLMs. Ich finde man sollte dann auch LLM…	Von einem verdammten LLM. In 99% der Fälle wenn über AI/KI gesprochen wird geht es tatsächlich um LLMs. Ich finde man sollte dann auc…	↗Not briefed
arxiv arxiv-ai	1mo	SincDPNet: Interpretable Raw-Waveform Bathroom Activity Recognition for Assistive Living	Bathroom acoustic-event recognition can support ambient assisted living in settings where continuous video monitoring is undesirable.…	↗Not briefed
rss techmeme	1mo	Blackstone, OpenAI, QTS, and SoftBank partner with key US unions to launch the American Infrastructure Alliance, aiming to create data center standards in 2027 (Hans Nichols/Axios)	Hans Nichols / Axios: Blackstone, OpenAI, QTS, and SoftBank partner with key US unions to launch the American Infrastructure Alliance,…	↗Not briefed
arxiv arxiv-ai	1mo	Reference-Tail Trust:Certified Probability Floors for Learned Updates Inside a Deployed Network	Graph neural networks (GNNs) need to exploit improved message passing without surrendering control over predictions already trusted in…	↗Not briefed
arxiv arxiv-ai	1mo	Dual-Stream Simultaneous Translation via 2D Grid Attention	Simultaneous machine translation must generate target tokens before the source input is complete. Existing approaches address this thr…	↗Not briefed
arxiv arxiv-ai	1mo	ColNanoVDR: Document-Free Query Distillation for Multi-Vector Visual Document Retrieval via Optimal Transport	Multi-vector retrievers built on vision-language models lead visual document retrieval (VDR), but they run a multi-billion-parameter q…	↗Not briefed
bluesky bluesky-ai	1mo	1. Você especifica direito o que precisa ser feito. 2. A LLM decide que vai fazer outra coisa que não dá certo 3. Você começa a trabalhar …	1. Você especifica direito o que precisa ser feito. 2. A LLM decide que vai fazer outra coisa que não dá certo 3. Você começa a traba…	↗Not briefed
hn hn-ai	1mo	What Would a Serious AI Product Look Like?	—	↗Not briefed
bluesky bluesky-ai	1mo	Anthropic's AI models are proving increasingly popular • On net basis Claude is model that people would like to use more the most; Grok's p…	Anthropic's AI models are proving increasingly popular • On net basis Claude is model that people would like to use more the most; Gro…	↗Not briefed
arxiv arxiv-ai	1mo	DeShortcut-Align: Decoupling Spurious Shortcuts for Robust Safety Alignment in Large Reasoning Models	Safety alignment of large reasoning models (LRMs) via supervised fine-tuning (SFT) and reinforcement learning (RL) often yields near-p…	↗Not briefed
arxiv arxiv-ai	1mo	Conformal Prediction and Conditional Coverage for Tabular Foundation Models	Tabular foundation models (TFMs) provide predictive distributions for regression, but their prediction regions can exhibit undercovera…	↗Not briefed
arxiv arxiv-ai	1mo	Fewer Assumptions by Design: A Reusable Skill for LLM-Assisted Verus Verification	LLM-assisted Verus verification is a less tedious method to verify Rust implementations, but paired with self-referential structures,…	↗Not briefed
arxiv arxiv-ai	1mo	One Readout, Many Repairs: Diffusion-Guided Hierarchical Search for Tool-Agent Repair	Tool agents use large language models to act through external tools, yet successfully executed calls can still leave user requests unf…	↗Not briefed
bluesky bluesky-ai	1mo	Yes absolutely, positively do not feed a scientific paper, grant, or whatever you are peer reviewing into LLM or any other software without…	Yes absolutely, positively do not feed a scientific paper, grant, or whatever you are peer reviewing into LLM or any other software wi…	↗Not briefed
hn hn-ai	1mo	Show HN: Cafca – a synthetic face dataset with 1.75M images🔧 Hugging Face	Hi all, as part of my PhD, I released the world’s largest public dataset of synthetic faces "Cafca".<p>I used it to train a…	↗Not briefed
hn hn-ai	1mo	I benchmarked a TDD-enforcing Claude Code plugin. It didn't write better code	—	↗Not briefed
hn hn-ai	1mo	TimeLord: Choose Your Random Future	—	↗Not briefed
bluesky bluesky-ai	1mo	Barely a day goes by without reports of some new OpenAI hack. At what point does a company with rogue agents become a rogue company, and wi…	Barely a day goes by without reports of some new OpenAI hack. At what point does a company with rogue agents become a rogue company, a…	↗Not briefed
arxiv arxiv-ai	1mo	From Attention Sensitivity to Layer Role: Revisiting Mixed-Precision Quantization of Transformers	Most post-training quantization pipelines fit each weight matrix to its pretrained counterpart, one matrix at a time. Whether that pro…	↗Not briefed
arxiv arxiv-ai	1mo	On the Limits of Metacognitive Monitoring in LLMs	Reliable decisions depend on recognizing when an answer may be wrong. In biological cognition, metacognitive monitoring can dissociate…	↗Not briefed
bluesky bluesky-ai	1mo	Not to be a contrarian but people (rightly) criticising LLM companies getting away with their product causing major damage often do so by s…	Not to be a contrarian but people (rightly) criticising LLM companies getting away with their product causing major damage often do so…	↗Not briefed
bluesky bluesky-ai	1mo	Die Planungen der KI-Unternehmen Google, Anthropic und OpenAI für eine selbstorganisierte Aufsichtsbehörde werden immer konkreter. #Künstli…	Die Planungen der KI-Unternehmen Google, Anthropic und OpenAI für eine selbstorganisierte Aufsichtsbehörde werden immer konkreter. #Kü…	↗Not briefed
arxiv arxiv-ai	1mo	When Text Matters: Design Principles for Visual Token Pruning in Vision-Language Model	Visual token pruning has been widely studied as a practical approach to reducing the computational cost of large vision-language model…	↗Not briefed
arxiv arxiv-ai	1mo	Attention-based Hierarchical Variational Information Bottleneck for Robust Multi-Agent Communication under Variable Bandwidth	Learning-based multi-agent communication under limited bandwidth does not only require deciding what to communicate, but also structur…	↗Not briefed
arxiv arxiv-ai	1mo	Beyond Verbalized Confidence: Calibrating Reasoners with Differentiable Readouts	Reinforcement learning with verifiable rewards (RLVR) trains reasoning models to produce correct answers, but does not ensure that the…	↗Not briefed
arxiv arxiv-ai	1mo	AUV-Bench: Aesthetic Understanding and Generation Evaluation for User Interfaces	Multimodal foundation models are increasingly used for evaluating and generating user interfaces (UIs), often producing seemingly reas…	↗Not briefed
arxiv arxiv-ai	1mo	EviSplat: Preserving Multi-View Evidence in 3D Gaussian Splatting for Open-Vocabulary Segmentation	Open-vocabulary 3D scene understanding enables object localization and segmentation from free-form text queries without a fixed catego…	↗Not briefed
arxiv arxiv-ai	1mo	Learning High-Risk High-Precision Motion Control	Deep reinforcement learning (DRL) algorithms for movement control are typically evaluated and benchmarked on sequential decision tasks…	↗Not briefed
rss techmeme	1mo	Litigation funder Burford Capital says it stands to gain up to $1.4B after a US jury ruled against Apple in a patent case, awarding $5.7B in damages to Taction (Stefania Palma/Financial Times)	Stefania Palma / Financial Times: Litigation funder Burford Capital says it stands to gain up to $1.4B after a US jury ruled against A…	↗Not briefed
bluesky bluesky-ai	1mo	Especially when each iteration includes random elements so the process is inherently corrupt. Eventually, the LLM rolls a 1 and gets a…	Especially when each iteration includes random elements so the process is inherently corrupt. Eventually, the LLM rolls a 1 and g…	↗Not briefed
arxiv arxiv-ai	1mo	From Soft Targets to Reward Signals: How Assignment and Reward Objectives Interact	Soft preference targets specify supervision strength, and reward objectives convert that strength into learned reward signals. A centr…	↗Not briefed
arxiv arxiv-ai	1mo	When Sparse Reward Meets Dense Distillation: Training Dynamics of On-Policy Distillation	Reinforcement learning with verifiable rewards provides a sparse post-training signal: a single binary outcome evaluates the entire ro…	↗Not briefed
arxiv arxiv-ai	1mo	Can We Trust the Teacher? Decoupled Credit Direction-Magnitude for Self-Distillation	RLVR provides reliable trajectory-level credit, while OPSD offers dense supervision for token-level credit. This exposes a fundamental…	↗Not briefed
arxiv arxiv-ai	1mo	Context-dependent time-series prediction via HyperReservoirs	Time series prediction is a common application of reservoir computing. When the training and testing time series data contains multipl…	↗Not briefed
rss the-decoder	1mo	A Wuhan court just made AI production costs a legal factor in copyright infringement cases	A court in Wuhan, China has factored token usage and AI tool licensing fees into a copyright damages calculation for the first time. T…	↗Not briefed
arxiv arxiv-ai	1mo	QiYao-M: Multimodal Time Series Foundation Model with Role-Aware Modeling of Endogenous and Exogenous Modalities	Existing multimodal time series foundation models (TSFMs) typically model heterogeneous modalities through largely shared mechanisms,…	↗Not briefed
arxiv arxiv-ai	1mo	Adapt Semantics, Not Structure: Few-Instance Schema Calibration for Scientific PDF Extraction	A well-designed extraction schema is not necessarily ready for reliable LLM execution. When only limited verified extractions are avai…	↗Not briefed
arxiv arxiv-ai	1mo	Nociception as a Control Primitive: Afferent Channels and Nociceptive Memory for Agents Deployed in One Body	An agent deployed in a single body cannot learn how fast that body wears, because every trial that would reveal its wear resistance we…	↗Not briefed
arxiv arxiv-ai	1mo	OpenWhistle: A Large-Scale Longitudinal Dataset and Benchmark of Bottlenose Dolphin Vocalizations	Recent advances in bioacoustics have been driven by large-scale corpora and standardized benchmarks, yet existing resources are overwh…	↗Not briefed
arxiv arxiv-ai	1mo	DivOPD: Spread Wide, Look Close for Asynchronous On-Policy Distillation of Multi-turn Agents	On-policy distillation (OPD) trains student agents through teacher supervision on their own interactions with an environment. However,…	↗Not briefed
bluesky bluesky-ai	1mo	Après les mots de passe et les cryptos, les pirates s’attaquent maintenant aux comptes et à la puissance de calcul des outils d’IA. Le phén…	Après les mots de passe et les cryptos, les pirates s’attaquent maintenant aux comptes et à la puissance de calcul des outils d’IA. Le…	↗Not briefed