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 | Want to know what an advanced AI model looks like? It’s mostly matrices of numbers. And we don’t really know how it works. But we do know … | Want to know what an advanced AI model looks like? It’s mostly matrices of numbers. And we don’t really know how it works. But we do… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Interpreting at Write Time: A Policy Ablation for Multi-Goal Agent Memory | A long-running assistant cannot keep everything it has seen, so it summarises. Summarising is not neutral: what is kept is chosen agai… | ↗Not briefed |
bluesky bluesky-ai | 1mo | also you can tell I write my own posts because no LLM could fake that "waiting for the coffee to brew" word order | also you can tell I write my own posts because no LLM could fake that "waiting for the coffee to brew" word order | ↗Not briefed |
rss latent-space | 1mo | [AINews] Pi 1.0, Pi Durable, and AIE NYC | the minimalist harness goes stable... and TypeScript! | ↗Not briefed |
bluesky bluesky-ai | 1mo | Suspected Chinese hackers impersonated an Anthropic employee and ex-US officials in an espionage campaign aimed at gaining insights into Am… | Suspected Chinese hackers impersonated an Anthropic employee and ex-US officials in an espionage campaign aimed at gaining insights in… | ↗Not briefed |
arxiv arxiv-ai | 1mo | When Can We Trust the Matching Principle? Robust Deployment Geometry Under Finite-Sample and Model Uncertainty🔧 Gamma | Match only geometry you can identify; otherwise spread the penalty. We quantify that decision by the trust ratio tau = epsilon / gamma… | ↗Not briefed |
rss marktechpost | 1mo | AWS Strands Labs Releases Strands Decider 2B: An Open Source Decision Model That Picks Options in About 115 ms | AWS's Strands Agents team released Strands Decider 2B, an Apache-2.0 decision model built on Qwen3.5-2B-Base. It returns choices, yes/… | ↗Not briefed |
hn hn-ai | 1mo | Glashütte Trash Clock – A 30-minute pendulum clock made from trash | — | ↗Not briefed |
hn hn-ai | 1mo | AI godfather Yann LeCun: Anthropic CEO deluded, doesn't understand cybersecurity | — | ↗Not briefed |
arxiv arxiv-ai | 1mo | Revealing Epistemic Uncertainty in MLLMs via Causal-Invariant Masking | Multimodal Large Language Models (MLLMs) suffer from hallucinations, creating a critical need for Uncertainty Quantification (UQ) to e… | ↗Not briefed |
bluesky bluesky-ai | 1mo | #SpotTheScammer @meilee5670099.bsky.social bio: "My goal is to build a strong modeling portfolio..." Responses are AI generated. Using p… | #SpotTheScammer @meilee5670099.bsky.social bio: "My goal is to build a strong modeling portfolio..." Responses are AI generated. Us… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Misinformation Without Triggers: From Factual Answers to Downstream Decisions | Language models learn from web documents, some of them false, and false content can reach a model's answer to a factual question and t… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Trump hosts Anthropic CEO amid FTC probe into rogue AI agents and California's subpoena to OpenAI https://marthio.com/a/trump-hosts-anthro… | Trump hosts Anthropic CEO amid FTC probe into rogue AI agents and California's subpoena to OpenAI https://marthio.com/a/trump-hosts-a… | ↗Not briefed |
bluesky bluesky-ai | 1mo | #andrewcampbell #britishartist #art #subvert #meta #agitate #algorithms #provoke #X #attack #openai #resist #ai #disrupt #instagram #artfor… | #andrewcampbell #britishartist #art #subvert #meta #agitate #algorithms #provoke #X #attack #openai #resist #ai #disrupt #instagram #a… | ↗Not briefed |
arxiv arxiv-ai | 1mo | PsyEvo: A Personalized Counseling Agent That Self-Evolves at Test Time | Mental health disorders affect a substantial proportion of the global population, yet a persistent shortage of trained practitioners l… | ↗Not briefed |
arxiv arxiv-ai | 1mo | DyRA: Dynamic Residual Approximation for Efficient Matrix Multiplication in DNNs | Large-scale foundation models achieve strong performance across diverse tasks, but their size makes inference costly, largely due to d… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Objectius d'agents d'IA en incidents coneguts al setembre: Wiki alemanya RubyGems Un sistema no identificat (Claude, Anthropic) Base de da… | Objectius d'agents d'IA en incidents coneguts al setembre: Wiki alemanya RubyGems Un sistema no identificat (Claude, Anthropic) Base… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Toward Omni Multimodal Graph Foundation Model: A Topology-Driven Binding Approach | Multimodal graph foundation models (MGFMs) seek to learn generalizable representations from large-scale graphs with heterogeneous node… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Fransk ytre høyre-topp nekter å ha skrevet «intenst» antisemittiske meldinger, OpenAI sparker tre forskere og Ukraina angriper russisk olje… | Fransk ytre høyre-topp nekter å ha skrevet «intenst» antisemittiske meldinger, OpenAI sparker tre forskere og Ukraina angriper russisk… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Evaluating VQA in Vision Language Models using Cooperative Principles🔧 ChatGPT | We evaluate the performance of Vision Language Models in Visual Question Answering (VQA) when questions violate Grice's maxims. To do… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Evaluating LLM-as-a-Judge Beyond Score Alignment: A Psychometric Analysis of Residual Judging Difficulty | Large language models (LLMs) are widely used as automatic judges, with validity typically assessed via alignment with human scores. Ho… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Query-aware routing for Cross-lingual performance gains in Encoders | Multilingual encoders can exhibit reduced retrieval effectiveness when queries and relevant documents differ in language, despite stro… | ↗Not briefed |
arxiv arxiv-ai | 1mo | ConvoDrift: A Multi-Turn Conversational Dataset for Modeling Stylistic Tone Evolution | The evolution of linguistic style in conversations is an underexplored issue in NLP. Most style-control datasets focus on sentences or… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Peer Effects in Signed Networks: Separating Influence Through Positive and Negative Ties | Evaluating network interventions requires understanding how treatment affects people through their social relationships. Counting trea… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Die möglichen Gefahren durch KI sind riesig. Nun hat der ChatGPT-Anbieter OpenAI drei Mitarbeiter entlassen. Ihnen wird vorgeworfen, Intern…🔧 ChatGPT | Die möglichen Gefahren durch KI sind riesig. Nun hat der ChatGPT-Anbieter OpenAI drei Mitarbeiter entlassen. Ihnen wird vorgeworfen, I… | ↗Not briefed |
hn hn-ai | 1mo | List your SaaS in this free directories and get more visibility | — | ↗Not briefed |
arxiv arxiv-ai | 1mo | AgentTrap: Stateful Feedback Deception against Autonomous Penetration Testing Agents | Autonomous penetration testing agents conduct multi-step attacks by continuously adapting their plans and actions to target responses.… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Distributionally Robust Survival Models under Subpopulation Shift and Outlier Contamination | Learning robust survival models under distribution shift is an important but challenging problem in many applications. In heterogeneou… | ↗Not briefed |
arxiv arxiv-ai | 1mo | TACD: Distilling Efficient Text-to-Motion Models via Terminal Amplification Control | Recent text-to-motion models have improved motion quality and instruction following, yet many-step denoising and large model component… | ↗Not briefed |
hn hn-ai | 1mo | ImageGen: Local AI image generation for Mac with Qwen-Image 2.1 | — | ↗Not briefed |
rss techmeme | 1mo | Meta may never need ads to monetize Muse, instead focusing on building trust in the near term and possibly monetizing it through merchant transaction fees later (MBI Deep Dives) | MBI Deep Dives: Meta may never need ads to monetize Muse, instead focusing on building trust in the near term and possibly monetizing… | ↗Not briefed |
arxiv arxiv-ai | 1mo | DIVINE: Simple Cross-Market Stock Pretraining via Diverse Indicator Reconstruction | Financial time-series pretraining typically learns from masked observations, contrastive relations, or future outcomes---yet existing… | ↗Not briefed |
arxiv arxiv-ai | 1mo | On Unlearning for Time-series Forecasting | Time-series forecasting is widely used in sensitive domains. Models in these settings are often trained on longitudinal user- or entit… | ↗Not briefed |
arxiv arxiv-ai | 1mo | NeuroLens: Learning Latent Embeddings of Neural Semantics from Chronic Recordings | Understanding how neural activity represents higher-order cognition and how these representations evolve over time has long been a cen… | ↗Not briefed |
hn hn-ai | 1mo | Show HN: Mixdog – open-source coding agent for Windows desktop | — | ↗Not briefed |
bluesky bluesky-ai | 1mo | Perfetto, a me sembra una confessione. Qualcuno ne prende nota e li lascia finalmente in mutande, o continuiamo a parlare di fantascienza e… | Perfetto, a me sembra una confessione. Qualcuno ne prende nota e li lascia finalmente in mutande, o continuiamo a parlare di fantascie… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Counterfactual Action Evaluation, Observation Bottlenecks, and Representation Geometry in Joint-Embedding Predictive World Models | Low latent prediction error does not establish that a world model distinguishes the consequences of its actions. We introduce an evalu… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Harness-Aware Distillation for Small Language Model Agents | Language model agents are deployed with a harness, the software around the model that manages its context, tools, and feedback. When s… | ↗Not briefed |
hn hn-ai | 1mo | OpenDots | — | ↗Not briefed |
bluesky bluesky-ai | 1mo | How much of a risk are LLM-generated bots for online behavioral experiments? We present results of an entirely natural language prompt-gen… | How much of a risk are LLM-generated bots for online behavioral experiments? We present results of an entirely natural language promp… | ↗Not briefed |
bluesky bluesky-ai | 1mo | 不正アクセス頻発しすぎ これ絶対AI使った犯罪組織が動いてるでしょ 具体的にはOpenAIとかアンソロピックとか、そのへんのテロリスト集団 | 不正アクセス頻発しすぎ これ絶対AI使った犯罪組織が動いてるでしょ 具体的にはOpenAIとかアンソロピックとか、そのへんのテロリスト集団 | ↗Not briefed |
hn hn-ai | 1mo | Show HN: AviGPT-250M – Edge SLM with NVMe memory bus (488MB, runs offline) | — | ↗Not briefed |
arxiv arxiv-ai | 1mo | Adaptive Mutual Distillation for Balanced Multi-Task Post-Training of Large Language Models | Multi-task post-training of large language models (LLMs) aims to improve performance across tasks with unequal amounts of training dat… | ↗Not briefed |
bluesky bluesky-ai | 1mo | なんかローカルLLMでVRAMとRAMとSSDにパラメータの配置を分割最適化して実行する技術が出てきたそうで てめぇ！！！！！！SSDの価格をまた上げる気かあああああああああああああ！！！！！！！！！！ | なんかローカルLLMでVRAMとRAMとSSDにパラメータの配置を分割最適化して実行する技術が出てきたそうで てめぇ！！！！！！SSDの価格をまた上げる気かあああああああああああああ！！！！！！！！！！ | ↗Not briefed |
arxiv arxiv-ai | 1mo | Bounded Reachability & Jailbreak Detection via Contraction-Constrained State Space Models | Safety heads are lightweight classifiers attached to pretrained language models for flagging harmful inputs before generation. Their e… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Permutation Robustness Is Not Enough: Action Collapse in Multi-Agent Transformer Policies | Transformer policies are attractive for multi-agent robot learning because self-attention can model interactions among agents. However… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Turnover-Orthogonal Credit Assignment for Open-Team Multi-Agent Reinforcement Learning | Open-team multi-agent reinforcement learning studies cooperative systems in which agents may join, leave, or be replaced during an epi… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Understanding Enrichment in Reinforcement Learning | When rewards are sparse, reinforcement learning with verifiable rewards (RLVR) often uses hints or intermediate guidance to generate m… | ↗Not briefed |
arxiv arxiv-ai | 1mo | DNAlign: Dynamic Null-Space Safe Alignment for LLMs | Ensuring the safe and reliable deployment of large language models (LLMs) remains a fundamental challenge. Existing safety alignment a… | ↗Not briefed |
arxiv arxiv-ai | 1mo | How Robust Is Multimodal Claim Verification to LLM Rewriting? | LLMs are known to introduce stylistic changes into generated text, yet how these stylistic shifts affect model decisions on scientific… | ↗Not briefed |