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 |
|---|---|---|---|---|
arxiv arxiv-ai | 2mo | (Whose defaults?) Is artificial intelligence reorienting archaeological methods? | Generative AI and the practice of "vibe coding" are changing how archaeologists carry out computational research, but their effects on… | ↗Not briefed |
bluesky bluesky-ai | 2mo | I think it’s maybe better to think of these LLM training cybersecurity incidents as superfund sites, AI companies dumping toxic waste in th… | I think it’s maybe better to think of these LLM training cybersecurity incidents as superfund sites, AI companies dumping toxic waste… | ↗Not briefed |
bluesky bluesky-ai | 2mo | Die CDU ist Anthropic weit voraus Mit Friedrich Merz hat sie die KI als Bundeskanzler installiert, die uns alle sehr viel schneller umbrin… | Die CDU ist Anthropic weit voraus Mit Friedrich Merz hat sie die KI als Bundeskanzler installiert, die uns alle sehr viel schneller u… | ↗Not briefed |
arxiv arxiv-ai | 2mo | FlexComp: One Model for Every Ratio in Context Compression | Soft context compression condenses a context into a few memory tokens that a frozen LLM consumes in place of the raw text, but existin… | ↗Not briefed |
bluesky bluesky-ai | 2mo | KI-Entwickler warnen vor der möglichen Auslöschung der Menschheit – und treiben genau diese Entwicklung gleichzeitig voran. Viel zynischer … | KI-Entwickler warnen vor der möglichen Auslöschung der Menschheit – und treiben genau diese Entwicklung gleichzeitig voran. Viel zynis… | ↗Not briefed |
arxiv arxiv-ai | 2mo | Agentic Share-of-Search: A Multi-Agent AI System for Competitive Decision-Making in LLM-Mediated E-Commerce | AI shopping assistants increasingly redirect consumer discovery, creating an urgent need for tools that support seller-side competitiv… | ↗Not briefed |
bluesky bluesky-ai | 2mo | Ļoti grūti domāt kategorijās "Latvija2050", "nākamais Nacionālās attīstības plāns", pat "nākamā Saeimas sasaukuma darāmie darbi", kad uz (t… | Ļoti grūti domāt kategorijās "Latvija2050", "nākamais Nacionālās attīstības plāns", pat "nākamā Saeimas sasaukuma darāmie darbi", kad… | ↗Not briefed |
bluesky bluesky-ai | 2mo | "your rant lands well" thanks LLM, I spent 11 minutes on it | "your rant lands well" thanks LLM, I spent 11 minutes on it | ↗Not briefed |
bluesky bluesky-ai | 2mo | So the tech broligarchs who have been filching our work are now claiming they’ve spawned an AI so monstrous it’s going to lead to our extin… | So the tech broligarchs who have been filching our work are now claiming they’ve spawned an AI so monstrous it’s going to lead to our… | ↗Not briefed |
hn hn-ai | 2mo | OpenAI: Defense Factory | — | ↗Not briefed |
arxiv arxiv-ai | 2mo | Can LLMs Follow Medical Expert Logic? A Benchmark for Hierarchical Logical Consistency in Risk-of-Bias Assessment | Evidence-based medicine demands strict logical consistency, yet current evaluations of large language models (LLMs) prioritize superfi… | ↗Not briefed |
arxiv arxiv-ai | 2mo | Exploring Second-Order Pattern Recognition in Speaker Recognition | In classical pattern recognition tasks, neural networks are trained to recognise human-defined patterns for model inputs. Some Explain… | ↗Not briefed |
bluesky bluesky-ai | 2mo | Opwinding in de wiskunde: OpenAI lost miljoendollarprobleem op – in 88 uur | Opwinding in de wiskunde: OpenAI lost miljoendollarprobleem op – in 88 uur | ↗Not briefed |
hn hn-ai | 2mo | Curated List of Osint Tools | — | ↗Not briefed |
hn hn-ai | 2mo | Kelivo: A Flutter LLM Chat Client. Support Mobile and Desktop | — | ↗Not briefed |
hn hn-ai | 2mo | GPS and the Limits of Control | — | ↗Not briefed |
arxiv arxiv-ai | 2mo | SemVerBench: Benchmarking LLM Comprehension of Version-Constraint Resolution Semantics🔧 GPT-5 | Large language model (LLM) coding agents constantly decide whether a version satisfies a constraint such as ^1.2.3 or >=2.0,<3, yet th… | ↗Not briefed |
bluesky bluesky-ai | 2mo | I resigned from Anthropic today. i was the assistant jizz mopper | I resigned from Anthropic today. i was the assistant jizz mopper | ↗Not briefed |
bluesky bluesky-ai | 2mo | DeepSeek v4.1 Flash is just another level 🤯 huggingface.co/deepseek-ai/... - Asymmetric Causal-Encoder-Decoder: 550B MoE, input 8B / ou… | DeepSeek v4.1 Flash is just another level 🤯 huggingface.co/deepseek-ai/... - Asymmetric Causal-Encoder-Decoder: 550B MoE, input 8B… | ↗Not briefed |
rss marktechpost | 2mo | DeepSeek AI Released DeepSeek-V4.1-Flash with 1M Context, FP4 KV Cache, and Cross-Layer Attention Reuse | Long-horizon agents have turned LLM serving into an input-heavy workload. Repeated prefills and million-token contexts leave KV caches… | ↗Not briefed |
bluesky bluesky-ai | 2mo | "Volgens de partij zijn de posts op het profiel van Jetten wel degelijk persoonlijk, ook al zijn ze door een AI-model geschreven. ‘Posts zi… | "Volgens de partij zijn de posts op het profiel van Jetten wel degelijk persoonlijk, ook al zijn ze door een AI-model geschreven. ‘Pos… | ↗Not briefed |
arxiv arxiv-ai | 2mo | Debate-to-Skill: Capability-Bound Process Supervision for Industrial Query-to-Agent Annotation | Industrial query-to-agent matching fails when topical relevance is mistaken for executable capability, especially on long-tail and bou… | ↗Not briefed |
bluesky bluesky-ai | 2mo | OpenAI used their creative path to solve the conjecture after the mathematicians had used ChatGPT. "There is a certain (frankly unacceptab…🔧 ChatGPT | OpenAI used their creative path to solve the conjecture after the mathematicians had used ChatGPT. "There is a certain (frankly unacc… | ↗Not briefed |
bluesky bluesky-ai | 2mo | 🚨🚨🚨 Seems like another Mathematician comes forward with concerns about OpenAI stealing unpublished math research and then publishing it … | 🚨🚨🚨 Seems like another Mathematician comes forward with concerns about OpenAI stealing unpublished math research and then publishin… | ↗Not briefed |
bluesky bluesky-ai | 2mo | Es propaganda para la que la prensa colabora en parte por ignorancia y en parte por intereses de los que tienen los medios. Esto decía Tim… | Es propaganda para la que la prensa colabora en parte por ignorancia y en parte por intereses de los que tienen los medios. Esto decí… | ↗Not briefed |
hn hn-ai | 2mo | Ex-Anthropic researcher Jacob Coxon says AI development poses risk to humans | — | ↗Not briefed |
arxiv arxiv-ai | 2mo | Hierarchical Clustering Can Jointly Satisfy Richness, Consistency, and Scale Invariance | Despite its ubiquity, clustering lacks a universally accepted definition of what is a cluster. Kleinberg's Impossibility Theorem forma… | ↗Not briefed |
arxiv arxiv-ai | 2mo | Beyond Visual Quality: Evaluating Physical Consistency under Ego-Motion with EgoGenEval | Recent visual generators produce high-fidelity images yet often violate physical consistency under ego-motion, limiting their use for… | ↗Not briefed |
arxiv arxiv-ai | 2mo | Breaking Predictions Is Not Enough: Specified-Foil Counterfactuals for Temporal Graphs | Temporal graph counterfactual explanations typically change past events to change or invalidate an original prediction, while leaving… | ↗Not briefed |
bluesky bluesky-ai | 2mo | This is what should be the push by critiques, force stronger regulation as unless you have a way of convincing the oligarchy capitalism is … | This is what should be the push by critiques, force stronger regulation as unless you have a way of convincing the oligarchy capitalis… | ↗Not briefed |
arxiv arxiv-ai | 2mo | Semi-Tensor Product-Based Multi-Term Randomized T-SVD and Its Visual Applications | Tensor singular value decomposition (T-SVD), which is built upon the tensor-tensor product (t-product), has emerged as a powerful tool… | ↗Not briefed |
bluesky bluesky-ai | 2mo | 最近、Claudeを使っていて気になるのは、勝ち負けを決めるディベートをしているのではないのに、わざと関係の無いデータや、うそ、ちょっとずれた事実をもってきて議論に「勝とう」とするところと、さらにそうした「うそ」について指摘すると、認めつつもムキになるところです。本人(人?)い… | 最近、Claudeを使っていて気になるのは、勝ち負けを決めるディベートをしているのではないのに、わざと関係の無いデータや、うそ、ちょっとずれた事実をもってきて議論に「勝とう」とするところと、さらにそうした「うそ」について指摘すると、認めつつもムキになるところです。本人… | ↗Not briefed |
bluesky bluesky-ai | 2mo | Paul Christiano, reconhecido expert sobre IA e segurança, ao anunciar seu ingresso no Conselho da OpenAI: | Paul Christiano, reconhecido expert sobre IA e segurança, ao anunciar seu ingresso no Conselho da OpenAI: | ↗Not briefed |
arxiv arxiv-ai | 2mo | When does a spectral prior help graph learning? Connectivity-loss estimation under road-network disruptions | Rapid evaluation of many simultaneous road-link disruptions requires a practical compromise between exact spectral recomputation and l… | ↗Not briefed |
bluesky bluesky-ai | 2mo | Le contexte. OpenAI vient d’annoncer avoir « résolu » Navier-Stokes, un des sept problèmes du millénaire, avec un essaim de 10 000 agents I… | Le contexte. OpenAI vient d’annoncer avoir « résolu » Navier-Stokes, un des sept problèmes du millénaire, avec un essaim de 10 000 age… | ↗Not briefed |
bluesky bluesky-ai | 2mo | Dollsys deepseek flash infra when | Dollsys deepseek flash infra when | ↗Not briefed |
arxiv arxiv-ai | 2mo | LILA: Calibration-Free Structured Pruning of Large Language Models via Latent Spectral Geometry | Structured pruning of large language models (LLMs) offers hardware-efficient compression, yet existing methods require calibration dat… | ↗Not briefed |
hn hn-ai | 2mo | How to make GPT-6 Astra creative | — | ↗Not briefed |
bluesky bluesky-ai | 2mo | new deepseek looks fantastic and i am not going to be able to sleep until they drop the paper so i can see what they did | new deepseek looks fantastic and i am not going to be able to sleep until they drop the paper so i can see what they did | ↗Not briefed |
bluesky bluesky-ai | 2mo | “L’AI in 10 anni potrà ucciderci tutti”. Viaggio nelle (tante) denunce dalla Silicon Valley Dopo la dimissioni di Coxon da Anthropic, racc… | “L’AI in 10 anni potrà ucciderci tutti”. Viaggio nelle (tante) denunce dalla Silicon Valley Dopo la dimissioni di Coxon da Anthropic,… | ↗Not briefed |
bluesky bluesky-ai | 2mo | Lo del ingeniero de Anthropic que renuncia y advierte que es peligroso mememe es PUBLICIDAD para que todos corramos en círculos y nos digam… | Lo del ingeniero de Anthropic que renuncia y advierte que es peligroso mememe es PUBLICIDAD para que todos corramos en círculos y nos… | ↗Not briefed |
arxiv arxiv-ai | 2mo | DRG-MAPPO: Hierarchical Dynamic Role-Graph Multi-Agent Reinforcement Learning for Cooperative Air Combat | Multi-Agent Reinforcement Learning (MARL) has emerged as a pivotal paradigm for complex decision-making in autonomous systems and air… | ↗Not briefed |
arxiv arxiv-ai | 2mo | terms.txt: A Consent and Compensation Protocol for Agentic Web Access | The open web ran on an unwritten bargain: sites admitted crawlers, and search engines sent visitors back. Public measurements show tha… | ↗Not briefed |
rss openai-news | 2mo | Expanding AI access and cyber defense for federal, state, local, and tribal governments | OpenAI and GSA will offer eligible federal, state, local, and tribal governments $0 license fees, 50% off usage, and expanded cyber de… | ↗Not briefed |
bluesky bluesky-ai | 2mo | I wonder how many of these people do in fact believe in some version of "eventually we will create a universally persuasive LLM" | I wonder how many of these people do in fact believe in some version of "eventually we will create a universally persuasive LLM" | ↗Not briefed |
bluesky bluesky-ai | 2mo | New theory: the openai image models are bad ONLY on food | New theory: the openai image models are bad ONLY on food | ↗Not briefed |
bluesky bluesky-ai | 2mo | Important to recognise distinction between machine learning and LLM-based "AI". ML has many useful applications in science/medicine & busin…🔧 ChatGPT | Important to recognise distinction between machine learning and LLM-based "AI". ML has many useful applications in science/medicine &… | ↗Not briefed |
bluesky bluesky-ai | 2mo | Membros das equipes de alinhamento (setor que monitora se sistemas de IA estão agindo de forma segura, previsível e ética) da Anthropic e O… | Membros das equipes de alinhamento (setor que monitora se sistemas de IA estão agindo de forma segura, previsível e ética) da Anthropi… | ↗Not briefed |
bluesky bluesky-ai | 2mo | I had to recheck. And yes. MY account is not allowed under the Ordoalea company rules to install Copilot, and my computer is flagged as "co… | I had to recheck. And yes. MY account is not allowed under the Ordoalea company rules to install Copilot, and my computer is flagged a… | ↗Not briefed |
arxiv arxiv-ai | 2mo | A Fragility Spectrum for Recursive Language-Model Training | Model-generated text is finding its way back into training corpora, and there is plenty of evidence that training on such data over an… | ↗Not briefed |