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 | 28d | I fear it's a series of arms races against AI all the way down for a declarative model of public service delivery. As a French journalist f… | I fear it's a series of arms races against AI all the way down for a declarative model of public service delivery. As a French journal… | ↗Not briefed |
arxiv arxiv-ai | 28d | Qwen-MT-Lumen at WMT2026: Instruction-Specialized Translation via Quality-Aware Training and On-Policy Distillation | We present Lumen, The open-source model is available at, our submission to the WMT 2026 General Machine Translation Shared Task. We fo… | ↗Not briefed |
bluesky bluesky-ai | 28d | A timely reminder of The Guardian's strategic partnership with OpenAI Stenographers for the worst people | A timely reminder of The Guardian's strategic partnership with OpenAI Stenographers for the worst people | ↗Not briefed |
bluesky bluesky-ai | 28d | My tumble dryer just hacked the Pentagon, lads. | My tumble dryer just hacked the Pentagon, lads. | ↗Not briefed |
bluesky bluesky-ai | 28d | Every press release from OpenAI and Anthropic in future looking like this... | Every press release from OpenAI and Anthropic in future looking like this... | ↗Not briefed |
arxiv arxiv-ai | 28d | Beyond Reachability: A Relation-Semantic Reasoning Framework for Biomedical Multi-hop QA | Biomedical multi-hop reasoning over knowledge graphs requires identifying a semantically valid relation sequence that supports the ans… | ↗Not briefed |
bluesky bluesky-ai | 28d | I see this in practice daily, but another question we must ask is how many of the complaints are legitimate? If LLM are enabling people who… | I see this in practice daily, but another question we must ask is how many of the complaints are legitimate? If LLM are enabling peopl… | ↗Not briefed |
hn hn-ai | 28d | OpenAI's dot assistant burned 300K+ tokens in a day without a single prompt | — | ↗Not briefed |
arxiv arxiv-ai | 28d | Discriminative Score Function: Turning Pretrained Models into Functional Generative Priors | Discriminative models are trained with task-specific objectives such as classification, detection, or representation learning, yet the… | ↗Not briefed |
arxiv arxiv-ai | 28d | Does the Linguistic Register of Thinking Matter? Wenyan as a Compact Reasoning Register for Chinese LLMs | Visible reasoning traces of Chinese reasoning-oriented LLMs are typically written in modern vernacular Chinese, which may be verbose u… | ↗Not briefed |
hn hn-ai | 28d | Mu-Silicium: community UEFI for ARM64 Consumer Devices | — | ↗Not briefed |
bluesky bluesky-ai | 28d | Die Huthi-Miliz im Jemen hat offenbar einen neuen Helfer bei der Entwicklung ihrer Raketen und Sprengköpfe entdeckt: künstliche Intelligenz… | Die Huthi-Miliz im Jemen hat offenbar einen neuen Helfer bei der Entwicklung ihrer Raketen und Sprengköpfe entdeckt: künstliche Intell… | ↗Not briefed |
bluesky bluesky-ai | 28d | Ein saudi-arabischer Ingenieur des ChatGPT-Entwicklers OpenAI soll ein Aufnahmegerät in einem Konferenzraum der Unternehmensführung install…🔧 ChatGPT | Ein saudi-arabischer Ingenieur des ChatGPT-Entwicklers OpenAI soll ein Aufnahmegerät in einem Konferenzraum der Unternehmensführung in… | ↗Not briefed |
arxiv arxiv-ai | 28d | ProtModernBERT: Compute-Efficient Protein Language Modeling | Large protein language models provide a powerful framework for learning sequence-function relationships and guiding experimental desig… | ↗Not briefed |
hn hn-ai | 28d | Lost in Implicit Routing: How LLM Agents Fail to Load Relevant Skills at Scale | — | ↗Not briefed |
arxiv arxiv-ai | 28d | MakSafe: A Spatiotemporal Semantic Framework for Formal Specification and Compliance Checking of Road Vehicle Safety Rules | Scenario-based validation is essential for automated road vehicles but cannot cover the full variability of openworld traffic. We pres… | ↗Not briefed |
bluesky bluesky-ai | 28d | As someone who works in the "breaking into devices" business I have to say that: 1. This is fear mongering, 2. LLMs generally act on input … | As someone who works in the "breaking into devices" business I have to say that: 1. This is fear mongering, 2. LLMs generally act on i… | ↗Not briefed |
arxiv arxiv-ai | 28d | Exact Cut Complexity of Equal-Length Proportional Cake Cutting | We study proportional cake cutting under an equal-length constraint, where each of $n$ agents must receive a bundle of length exactly… | ↗Not briefed |
hn hn-ai | 28d | Show HN: Doom running inside a JavaScript datagrid | — | ↗Not briefed |
arxiv arxiv-ai | 28d | What Happens to the Goal? Distinct Failures of Policy Alignment in Goal Misgeneralization | Goal misgeneralization is a common failure mode in which an agent behaves competently out of distribution but does not pursue the inte… | ↗Not briefed |
bluesky bluesky-ai | 28d | How do I convince my parents that gen ai is literally unethical?? They want me to make a business out of "digital prints" using Claude or s… | How do I convince my parents that gen ai is literally unethical?? They want me to make a business out of "digital prints" using Claude… | ↗Not briefed |
rss marktechpost | 28d | Liquid AI Releases LFM2.5-24B-A2B: A 2B-Active MoE for On-Device Agents | Liquid AI released LFM2.5-24B-A2B, an open-weight 24B Mixture-of-Experts reasoning model with 2B active parameters. It targets local a… | ↗Not briefed |
arxiv arxiv-ai | 28d | Mitigating Video Hallucinations via Attention-Guided Representation Perturbation | Video large language models (Video LLMs) often generate responses unsupported by the input video. Contrastive decoding (CD) can mitiga… | ↗Not briefed |
rss bing-news-ai | 28d | Not Nvidia. Not Micron. This Under-the-Radar Artificial Intelligence (AI) Chip Stock Could Be the Biggest Winner in 2027. | Missed Nvidia in 2009? This Rare Signal Is Flashing Again. In 2009, a "Double Down" signal flashed for a little-known chipmaker called… | ↗Not briefed |
arxiv arxiv-ai | 28d | TIMERSI: Towards Self-Improved Agentic Deployment for Time-Series Foundation Models | Time-series foundation models provide strong zero-shot forecast, but deployment for real- world applications still requires task-speci… | ↗Not briefed |
arxiv arxiv-ai | 28d | Recognizing Dialogue Ambiguity via Candidate Set Visual Grounding | Interpreting an ambiguous utterance in real-world interactions often requires context beyond a single utterance. When the utterance re… | ↗Not briefed |
arxiv arxiv-ai | 28d | ONCship: A Large-Scale Underwater Acoustic Corpus for Ship Recognition | Recent advances in deep learning have significantly improved underwater acoustic vessel recognition. However, existing publicly availa… | ↗Not briefed |
arxiv arxiv-ai | 28d | Articulation Upper Bounds for Language Models | Language captures only a subset of the perceptions, intuitions and intents that form meaning between speakers. What, then, are the lim… | ↗Not briefed |
bluesky bluesky-ai | 28d | Every new app starts with the same 20 screens. What if your data model built them for you? Underpeaks is a free, self-hosted CMS that gene… | Every new app starts with the same 20 screens. What if your data model built them for you? Underpeaks is a free, self-hosted CMS that… | ↗Not briefed |
bluesky bluesky-ai | 28d | In fact, I'd love to see every tech exec who approves multi-billion opex or capex on LLMs to have to *bet their job* on them producing at l… | In fact, I'd love to see every tech exec who approves multi-billion opex or capex on LLMs to have to *bet their job* on them producing… | ↗Not briefed |
hn hn-ai | 28d | OpenAI pursues acquisition of Trail of Bits security firm | — | ↗Not briefed |
arxiv arxiv-ai | 28d | Learning Individual Scientific Taste with Privileged Hindsight Distillation | Scientific taste shapes which research directions researchers pursue, and it varies across individuals. Existing work evaluates resear… | ↗Not briefed |
bluesky bluesky-ai | 28d | The full vulnerability report for the nastly #curl UAF problem is disclosed. It was submitted by people at Anthropic. The reproducer is a m… | The full vulnerability report for the nastly #curl UAF problem is disclosed. It was submitted by people at Anthropic. The reproducer i… | ↗Not briefed |
arxiv arxiv-ai | 28d | Uncertainty-Driven Self-Play for Robust Multi-Agent Driving Policies | Learning robust driving policies that generalize beyond their training domain remains a central challenge in motion planning. We propo… | ↗Not briefed |
hn hn-ai | 28d | Ace-Step 1.5 Open-Source Music Generation Model | — | ↗Not briefed |
arxiv arxiv-ai | 28d | New Model, Old Habits: Behavior Transfer under Model Swapping in Multi-Session LLM Agents | Multi-session agents can switch between LLMs to balance performance and inference cost. Such switches require transferring a history f… | ↗Not briefed |
arxiv arxiv-ai | 28d | Beyond State Capacity: Why Retrieval Breaks in Mamba | Mamba models underperform Transformers on in-context retrieval. This gap is usually attributed to Mamba's fixed-size recurrent state,… | ↗Not briefed |
arxiv arxiv-ai | 28d | SceneAE: Scene Generation and Reconstruction in Queryable 3D-Anchored Latent Space | Recent video world models have increasingly adopted 3D memory mechanisms to integrate historical observations into a lightweight and p… | ↗Not briefed |
hn hn-ai | 29d | Show HN: Recreating Gemini Argon 3D demo using ExtrudeAI and FreeCAD | — | ↗Not briefed |
arxiv arxiv-ai | 29d | RouteWeaver: Weaving Mode Selection and Execution into Unified LLM Routing | LLM routing dynamically selects and coordinates heterogeneous language models across diverse tasks. However, existing approaches are l… | ↗Not briefed |
arxiv arxiv-ai | 29d | BANGLAPLAN: A Benchmark for Constraint-Based Planning and Decision-Making in Bangla | Large language models (LLMs) increasingly help users arrange appointments, plan journeys, and allocate limited resources. These tasks… | ↗Not briefed |
hn hn-ai | 29d | Show HN: Ruleherd – audits agent instruction files for contradictions and drift | — | ↗Not briefed |
rss bing-news-ai | 29d | The billion-dollar defense AI revolution is here: Autonomous mission intelligence set to transform modern warfare | Billions in Defense Modernization Are Driving Demand for Advanced AI, Autonomous Coordination, and Next-Generation Battlefield Intelli… | ↗Not briefed |
bluesky bluesky-ai | 29d | has anthropic maybe considered that most people dont have 15 smart tvs in their home and hence they would have to distribute inference over… | has anthropic maybe considered that most people dont have 15 smart tvs in their home and hence they would have to distribute inference… | ↗Not briefed |
bluesky bluesky-ai | 29d | Is AI eating neoliberalism? www.taxresearch.org.uk/Blog/2026/10... | Is AI eating neoliberalism? www.taxresearch.org.uk/Blog/2026/10... | ↗Not briefed |
arxiv arxiv-ai | 29d | Runtime Monitoring of Supermartingale Certificates | Supermartingale certificates formally guarantee temporal properties of discrete and continuous-time stochastic dynamical systems. Howe… | ↗Not briefed |
arxiv arxiv-ai | 29d | ReGRA: Residual-Gradient Adaptation for Complementary Low-Rank Fine-Tuning | Low-rank adaptation (LoRA) enables parameter-efficient fine-tuning by learning low-rank weight updates. Increasing rank expands adapta… | ↗Not briefed |
arxiv arxiv-ai | 29d | RuleTrek: Benchmarking Agentic Rule Reasoning in Complex Scenarios | Large language models (LLMs) increasingly tackle complex real-world tasks in specialized domains. However, identifying and correctly a… | ↗Not briefed |
hn hn-ai | 29d | A perceptual color space for image processing (2020) | — | ↗Not briefed |
arxiv arxiv-ai | 29d | Share the Symbol, Not the Vector: Training-Free Fact Communication Across Language Models | Multi-agent systems allow language models to solve tasks collaboratively by sharing complementary information. These systems typically… | ↗Not briefed |