News
The raw AI news stream — everything we ingest across RSS, Hacker News, arXiv, GitHub, Bluesky, YouTube and more, newest first.
| Source | Age | Title | Summary | Status |
|---|---|---|---|---|
bluesky bluesky-ai | 1mo | J'y pense depuis que l'"affaire" est sortie. Et certains se posaient la question de qui sera le premier à revendiquer avoir écrit un bouqui… | J'y pense depuis que l'"affaire" est sortie. Et certains se posaient la question de qui sera le premier à revendiquer avoir écrit un b… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Beyond Sub-Gaussian Detector Scores: Robust Weighted Profile-Loss Change Point Detection for Human-LLM Text Segmentation | Mixed human-LLM documents require locating authorship transitions from detector scores whose reliability varies across text units. Exi… | ↗Not briefed |
hn hn-ai | 1mo | Stepper firmware for the Uno that stops the motor when your program stops | — | ↗Not briefed |
arxiv arxiv-ai | 1mo | WEFT: Scaling Tool-Use Post-Training for General-Purpose Agents | Recent efforts to scale tool-use post-training have largely centered on the synthesis of executable environments, which constitute onl… | ↗Not briefed |
arxiv arxiv-ai | 1mo | RNA Design via Conditioned Flow Matching and Finite-Policy Reinforcement Learning | RNA design aims to identify sequences that fold into specified secondary structures. Existing methods formulate the task as target-spe… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Anthropic: "AI may pose 'existential risks to humanity'" World governments: "Extinction is such a hot market right now, and we want a fat … | Anthropic: "AI may pose 'existential risks to humanity'" World governments: "Extinction is such a hot market right now, and we want a… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Architecture Alignment With Sparse Priors in Tabular Foundation Models | Tabular foundation models (TFMs) are increasingly popular because they deliver strong predictions on new datasets through in-context l… | ↗Not briefed |
arxiv arxiv-ai | 1mo | What You Observe Determines How You Identify Causal Effects: Evaluating Causal Models across Observational Views | Causal foundation models (CFMs) pre-trained on data generated from various structural causal models (SCMs) have been proposed for esti… | ↗Not briefed |
arxiv arxiv-ai | 1mo | SKILLLITE: Evidence-Guided Malicious Skill Auditing with Compact LLMs | As LLM-based agents perform increasingly complex tasks, Agent Skills have emerged as a flexible mechanism for extending their capabili… | ↗Not briefed |
bluesky bluesky-ai | 1mo | if sephiroth were real he would not allow there to be an anthropic | if sephiroth were real he would not allow there to be an anthropic | ↗Not briefed |
bluesky bluesky-ai | 1mo | The model training is facing extreme diminishing returns anyway, and the amount of AI generated content being fed back into training is hav… | The model training is facing extreme diminishing returns anyway, and the amount of AI generated content being fed back into training i… | ↗Not briefed |
bluesky bluesky-ai | 1mo | NOW THE PROBLEM: being boeing or airbus does not yield the multiples on revenue they want from their IPO. so the question is, how much can … | NOW THE PROBLEM: being boeing or airbus does not yield the multiples on revenue they want from their IPO. so the question is, how much… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Seeing Time: Visual-Temporal Representation Learning for Interpretable Time Series Clustering | Multivariate Time Series (MTS) clustering is an important tool in temporal data mining, aiming to discover latent group structures fro… | ↗Not briefed |
hn hn-ai | 1mo | Denial of Spend: New category of resource exhaustion attacks | — | ↗Not briefed |
bluesky bluesky-ai | 1mo | I honestly think #AI companies will be junk shares in a few years time. Their business model is unsustainable, based on theft, and most peo… | I honestly think #AI companies will be junk shares in a few years time. Their business model is unsustainable, based on theft, and mos… | ↗Not briefed |
hn hn-ai | 1mo | OpenAI: Tomorrow we are re-opening the Pro $200 subscription | — | ↗Not briefed |
arxiv arxiv-ai | 1mo | State Trace Rationale As Auxiliary Task in Reinforcement Learning | We propose STRAT, an auxiliary task that trains deep reinforcement learning (RL) agents to predict a short textual trace of their own… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Where the Model Changes Its Mind: Hindsight-Divergence Localization for Efficient Reinforcement Learning with Verifiable Rewards | Group-relative methods for reinforcement learning with verifiable rewards (RLVR) learn from differences in rollout outcomes. Independe… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Safer Content or Firmer Refusals? A Hybrid Perturbation Defense for Alignment under Harmful Fine-tuning | Fine-tuning-as-a-service lets users adapt a safety-aligned language model to their own data, but it also creates a harmful fine-tuning… | ↗Not briefed |
hn hn-ai | 1mo | US Forces Exit Iraq | — | ↗Not briefed |
arxiv arxiv-ai | 1mo | IronLLM: Forging Compact Edge-Native Language Models for Real-Time Embodied Intelligence | We present IronLLM-0.6B, a 654M-parameter language model designed for efficient on-device inference. IronLLM-0.6B combines a hybrid at… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Markovian Nonconvex ADMM for Reinforcement Learning: Bellman-Resolvent Stability Beyond Smooth Blocks | We identify and study a structural mechanism for Markovian nonconvex ADMM in reinforcement learning. Using finite discounted MDPs as a… | ↗Not briefed |
arxiv arxiv-ai | 1mo | When Upstream Messages Override Correct Answers: A Controlled Study of Multi-Agent LLM Collaboration | Multi-agent LLM systems rely on message passing among specialized agents to accomplish complex tasks. However, an upstream agent may p… | ↗Not briefed |
bluesky bluesky-ai | 1mo | AI Regulation Made Simple: ‘If you develop an AI Model/agent/whatever that ever hurts one single person, you, everyone who works at your c… | AI Regulation Made Simple: ‘If you develop an AI Model/agent/whatever that ever hurts one single person, you, everyone who works at y… | ↗Not briefed |
bluesky bluesky-ai | 1mo | AI時代にはこれまで以上に「顔の良さ」が重要になる可能性がある、AIは人間以上に顔の偏見が強いため https://gigazine.net/news/20260929-ai-use-facial-features-criminality-performance/ | AI時代にはこれまで以上に「顔の良さ」が重要になる可能性がある、AIは人間以上に顔の偏見が強いため https://gigazine.net/news/20260929-ai-use-facial-features-criminality-performance/ | ↗Not briefed |
arxiv arxiv-ai | 1mo | Rethinking Multimodal Fake News Detection in the Generative AI Era | Generative content is increasingly entering the production and dissemination of news, transforming fake news from manually fabricated… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Does the Unsafe Gradient Survive a Conversation? On the Fragility of Gradient-Based Jailbreak Detection in Multi-Turn Dialogue | Safety-aligned language models are commonly deployed as multi-turn assistants, which lets adversaries spread unsafe intent across seve… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Is the business profitable? No It’s on the road to profitability though, right? Also no But it will be extremely beneficial for the futur… | Is the business profitable? No It’s on the road to profitability though, right? Also no But it will be extremely beneficial for the… | ↗Not briefed |
arxiv arxiv-ai | 1mo | An Effective, Reliable, and Robust Framework for Human Activity Recognition Using Wearable Sensors | Human Activity Recognition (HAR) through wearable sensors greatly improves the quality of human life through its multiple applications… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Automated Screw Planning for Reduced Pelvic Fractures Based on Statistical Shape Models and Deep Learning | Percutaneous iliosacral screw fixation is an important minimally invasive treatment for unstable pelvic fractures. Because the sacroil… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Vanaf wanneer gaan we dit een hackerscollectief noemen? nos.nl/artikel/2632... | Vanaf wanneer gaan we dit een hackerscollectief noemen? nos.nl/artikel/2632... | ↗Not briefed |
arxiv arxiv-ai | 1mo | DSWM: Decomposed Spatio-Temporal World Model for Demand-Driven UAV Base Station Repositioning | Uncrewed aerial vehicle base stations (UAV-BSs) are expected to cover traffic demand that shifts across space and time, yet most repos… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Anthropic warns that Capex build out is unsustainable. Lol, jk its fine | Anthropic warns that Capex build out is unsustainable. Lol, jk its fine | ↗Not briefed |
hn hn-ai | 1mo | I built a small workflow execution engine around a deliberately narrow problem | — | ↗Not briefed |
bluesky bluesky-ai | 1mo | Study: How Keywords in Business Names Currently Impact Google Local Pack Rank and AI Search - whitespark.ca/blog/study-h... #seo #geo #llm … | Study: How Keywords in Business Names Currently Impact Google Local Pack Rank and AI Search - whitespark.ca/blog/study-h... #seo #geo… | ↗Not briefed |
arxiv arxiv-ai | 1mo | RolloutFaith: Auditing Persistent Internal Interventions in Visual World Model | Interpretability methods such as probes, activation patches and learned editors are designed to reveal or modify a model's current com… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Cripes, 80 pages of the prospectus is "be afraid, be very afraid", almost a third of the document! But is that more frightening for inves… | Cripes, 80 pages of the prospectus is "be afraid, be very afraid", almost a third of the document! But is that more frightening for… | ↗Not briefed |
hn hn-ai | 1mo | RuneBench: Agent Benchmark on RuneScape Gameplay Tasks | — | ↗Not briefed |
hn hn-ai | 1mo | Towards safety cases for frontier AI training | — | ↗Not briefed |
arxiv arxiv-ai | 1mo | On-Policy Visual Evidence Distillation | Visual agents solve problems by interleaving reasoning with image operations, and on-policy distillation (OPD) provides guidance from… | ↗Not briefed |
hn hn-ai | 1mo | As AI models go rogue, do you still trust OpenAI and Anthropic to stop them? | — | ↗Not briefed |
arxiv arxiv-ai | 1mo | You Cannot Recover What Was Never Measured: Quantifying the Information Ceiling of Ultra-Low-Field MRI Super-Resolution | Generative super-resolution models can turn portable 64 mT MRI into images that look like 3T scans, and the field evaluates them with… | ↗Not briefed |
arxiv arxiv-ai | 1mo | ARC-KV: Amortizing Anchor Search for Reconstruction-Based KV Cache Compaction | Long-context large language model inference is bottlenecked by KV caches that grow linearly with sequence length. This burden is espec… | ↗Not briefed |
bluesky bluesky-ai | 1mo | return of the king anthropic lol | return of the king anthropic lol | ↗Not briefed |
bluesky bluesky-ai | 1mo | "Anthropic outlined more prosaic risks, including the extreme concentration of its customer base, with close to a quarter of revenue last y… | "Anthropic outlined more prosaic risks, including the extreme concentration of its customer base, with close to a quarter of revenue l… | ↗Not briefed |
hn hn-ai | 1mo | Automating eval design and hillclimbing with Claude | — | ↗Not briefed |
bluesky bluesky-ai | 1mo | Within the next year the interest on AI spending will come due and likely OpenAI still won’t get the $1.5tn IPO it needs to pay off the Ven… | Within the next year the interest on AI spending will come due and likely OpenAI still won’t get the $1.5tn IPO it needs to pay off th… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Where Does Staleness Accumulate? Pool Aware Effective Staleness Control for Asynchronous RL in LLM Post-Training | Fully asynchronous reinforcement learning (RL) improves resource utilization in large language model post-training by overlapping roll… | ↗Not briefed |
arxiv arxiv-ai | 1mo | The Default Trap: Rethinking Plan Evaluation in Tool-Using LLM Agents | An executor can respond strongly to a change in a supplied plan's priority while showing a small change in the same information-select… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Calibrate the Decisions That Change the Future: On-Policy Post-Training Quantization for Multimodal Large Language Models🔧 Modal | Post-training quantization (PTQ) lowers deployment cost for multimodal large language models, but calibration typically reconstructs f… | ↗Not briefed |