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 | 25d | Hy Memory 🧠 New memory model for AI agents from Tencent Hunyuan huggingface.co/tencent/Hy-M... - 30B MoE / 3B active - Apache 2.0 - E… | Hy Memory 🧠 New memory model for AI agents from Tencent Hunyuan huggingface.co/tencent/Hy-M... - 30B MoE / 3B active - Apache 2.0… | ↗Not briefed |
arxiv arxiv-ai | 25d | Not All Test-Time Compute Is Evidence: Model-Specific Bias in LLM Reasoning | Test-time scaling methods implicitly assume that additional computation from the same model provides independent evidence. We show tha… | ↗Not briefed |
bluesky bluesky-ai | 25d | New blog post about Ellf and what's currently missing for NLP and code + data projects with coding agents. If we can make coding assistant… | New blog post about Ellf and what's currently missing for NLP and code + data projects with coding agents. If we can make coding assi… | ↗Not briefed |
arxiv arxiv-ai | 25d | MAS-Bai: Linguistically Informed Adaptive Multi-Agent Translation for Bowen-Script Bai | Large Language Models (LLMs) remain unreliable for extremely low-resource languages whose writing systems conflict with high-resource… | ↗Not briefed |
hn hn-ai | 25d | Show HN: A music video made end to end by Claude Code, full production repo | — | ↗Not briefed |
arxiv arxiv-ai | 25d | Non-Parametric Optimization for Scalable Learning in Stochastic Decision Problems | Stochastic optimization (SO) plays a central role in decision-making under uncertainty. Among SO problems, time-varying stochastic opt… | ↗Not briefed |
arxiv arxiv-ai | 25d | Validating LLM-generated political networks with survey-reported partisan exposure | Large language models (LLMs) are increasingly used to simulate human behavior in social science research, including decisions, interac… | ↗Not briefed |
hn hn-ai | 25d | I turned my brain MRI into an interactive 3D model you can explore | — | ↗Not briefed |
arxiv arxiv-ai | 25d | Framing Spatial Hearing within the Free Energy Principle: A Hierarchical Model of Perceptual Learning | Spatial hearing relies on dynamic internal representations to decode spatial information from ever-changing acoustic conditions. A fun… | ↗Not briefed |
arxiv arxiv-ai | 25d | FuzzInsight: Navigating the Efficiency-Effectiveness Trade-off in Coverage-Guided DNN Fuzzing | Collecting and labeling diverse test datasets for evaluating Deep Neural Networks (DNNs) is costly and time-consuming. This has motiva… | ↗Not briefed |
arxiv arxiv-ai | 25d | OfficeTown: Generative Organizations for Evolving Agent Intelligence | AI agents increasingly use files and software tools to produce professional deliverables. Such work draws on materials, decisions, and… | ↗Not briefed |
hn hn-ai | 25d | Show HN: Three philosophers vote on your questions | I'm RN developer in South Korea, i built this on my own. Non-native speaker, sorry for my English.<p>- Sometimes it is very hard… | ↗Not briefed |
arxiv arxiv-ai | 25d | Proportionality in Ranking Compression | Motivated by applications in reinforcement learning from human feedback, we consider a setting where a given collection of rankings ne… | ↗Not briefed |
hn hn-ai | 25d | Order to limit Starshield for Ukraine came from inside White House, sources say | — | ↗Not briefed |
arxiv arxiv-ai | 25d | LangNCO: Compiling Natural-Language Constraints into Inference-Time Interventions for Neural Combinatorial Optimization | Neural combinatorial optimization (NCO) rapidly produces high-quality solutions for problem structures encountered during training, bu… | ↗Not briefed |
hn hn-ai | 25d | Creatine may help shield the brain from damage linked to gut inflammation | — | ↗Not briefed |
arxiv arxiv-ai | 25d | A Stronger Teacher Within: Strengthening MoE Models via On-Policy Bidirectional Distillation | Reinforcement Learning with Verifiable Rewards (RLVR) has substantially improved Mixture-of-Experts (MoE) LLMs, but its outcome-level… | ↗Not briefed |
bluesky bluesky-ai | 25d | お声がけいただいたのでAnthropicのイベントにきた。ライゾマ真鍋さんとかの展示とトークセッション。 casabrutus.com/categories/d... | お声がけいただいたのでAnthropicのイベントにきた。ライゾマ真鍋さんとかの展示とトークセッション。 casabrutus.com/categories/d... | ↗Not briefed |
arxiv arxiv-ai | 25d | AMEND: Constraint-Guided Multi-Agent Text Detoxification with Isolated Critics | Text detoxification requires balancing toxicity removal, semantic preservation, and fluency. Existing methods typically optimize an un… | ↗Not briefed |
arxiv arxiv-ai | 25d | What LLMs Do in the Shadows: Reading Context Reliance from the Unembedding Dark Space | Detecting faithfulness hallucinations in large language models (LLMs) is crucial for their safe deployment in high-stakes domains. Rec… | ↗Not briefed |
bluesky bluesky-ai | 25d | Same reason I’m against llm coding: we built up good practices and principles to protect ourselves and build projects in a way that is mean… | Same reason I’m against llm coding: we built up good practices and principles to protect ourselves and build projects in a way that is… | ↗Not briefed |
bluesky bluesky-ai | 25d | The frightening thing is that using an LLM absolutely does not for you from the burden of having to think, it just kind of feels like it mi… | The frightening thing is that using an LLM absolutely does not for you from the burden of having to think, it just kind of feels like… | ↗Not briefed |
arxiv arxiv-ai | 25d | Tree-Aware Posterior Distillation for Long-Horizon Agentic Reinforcement Learning | Reinforcement learning with verifiable rewards has become a scalable post-training paradigm for large language models. However, traini… | ↗Not briefed |
bluesky bluesky-ai | 25d | そんなわけですちーぶんのMKAL再スタートして、clue1のセクション1が終わった！アポロチョコだよ！ 2度目だから前よりは早いか。糸は、切らずに渡してる。1度目に「こうすればよかった」と思ったところを丁寧に拾いながら、きちんと編んでいきたいと思ってる。 あと朝になったらAI… | そんなわけですちーぶんのMKAL再スタートして、clue1のセクション1が終わった！アポロチョコだよ！ 2度目だから前よりは早いか。糸は、切らずに渡してる。1度目に「こうすればよかった」と思ったところを丁寧に拾いながら、きちんと編んでいきたいと思ってる。 あと朝にな… | ↗Not briefed |
arxiv arxiv-ai | 25d | Do Speech BCIs Need Larger Models? Rethinking Neural Decoding beyond Scaling🔧 Modal | Recent speech brain-computer interfaces (BCIs) increasingly rely on large models and complex multi-stage pipelines for neural decoding… | ↗Not briefed |
hn hn-ai | 25d | Ask HN: Has anyone started using OpenAI Agents API | If you are building your agents, or just out of curiosity, have you tried the OpenAI Agents API yet? | ↗Not briefed |
bluesky bluesky-ai | 25d | can we admire how cute deepseek is whenever it REs things | can we admire how cute deepseek is whenever it REs things | ↗Not briefed |
arxiv arxiv-ai | 25d | Structural Affinity: An Algorithm for Coherent Cage-Based Shape Design | Cage-based deformation (CBD) parametrises 3D geometry through a small set of control handles, effectively enabling a variety of tasks… | ↗Not briefed |
hn hn-ai | 25d | Japan mistakenly issues warning for magnitude 9.9 earthquake | — | ↗Not briefed |
arxiv arxiv-ai | 25d | Chinese Decision: A Family of Language Models for Single-Pass Probabilistic Decision-Making | Many language-based applications require probabilities over a finite set of decisions rather than open-ended responses. We present Chi… | ↗Not briefed |
arxiv arxiv-ai | 25d | LongForge: Measuring and Improving Long-Context Instruction Data via Evidence-Assertion Reasoning Graphs | Existing quality control for long-context instruction-tuning data is a one-shot judgment inside the generation pipeline that leaves on… | ↗Not briefed |
arxiv arxiv-ai | 25d | Explaining Chess and Go in Natural Language | Modern chess engines play far beyond human ability, often choosing moves whose underlying ideas are difficult for humans to understand… | ↗Not briefed |
hn hn-ai | 25d | Unihertz won't release the GPL'd kernel source, so I ported LineageOS anyway | — | ↗Not briefed |
arxiv arxiv-ai | 25d | Algorithm Modelization: Transforming Optimization Algorithms into Evolvable Intelligence | Can algorithms be learned and evolved as models, rather than repeatedly redesigned as programs? We explore this question by introducin… | ↗Not briefed |
arxiv arxiv-ai | 25d | Nearly Linear Private Learning of Littlestone Classes | Differentially private learning aims to learn from data while protecting individual training examples. For binary classification, fini… | ↗Not briefed |
hn hn-ai | 25d | Ask HN: What world do you want to live in? | Everyone on here is quite familiar with the tech tool potential and the dystopian use trends. In some large part, these are social dec… | ↗Not briefed |
hn hn-ai | 25d | I recreated the Vision Pro palm-flip gesture on Apple Watch | — | ↗Not briefed |
hn hn-ai | 25d | Streaming a coding agent's UI in Flutter: what broke and how we fixed it | — | ↗Not briefed |
arxiv arxiv-ai | 25d | Cross-Centaur: Modeling Individual Decision-Making Across Psychological Tasks with Language Models | Large language models (LLMs) are increasingly used as computational models of human behavior. Recent models such as Centaur have advan… | ↗Not briefed |
arxiv arxiv-ai | 25d | QuAD: Quantitative Assessment via Distribution-Alignment for Qualitative Data Quality Enabled by a Content-Aware Multi-Agent System | Assessing whether qualitative responses collected from semi-structured questions are ready for a particular analysis goal is a persist… | ↗Not briefed |
bluesky bluesky-ai | 25d | I think the one valid use for AI would be creating a model that figures out how to adaptively block every new permutation of "add X to a mo… | I think the one valid use for AI would be creating a model that figures out how to adaptively block every new permutation of "add X to… | ↗Not briefed |
hn hn-ai | 25d | LibreNT – how AI slop infects Windows clean-room engineering | — | ↗Not briefed |
bluesky bluesky-ai | 25d | "Releasing over 700 files at once is not a demonstration of scholarship, but a demonstration of power. We urge mathematicians to discontinu… | "Releasing over 700 files at once is not a demonstration of scholarship, but a demonstration of power. We urge mathematicians to disco… | ↗Not briefed |
arxiv arxiv-ai | 25d | CIIR - Class-Indexed Interface Representation Learning | Multimodal learning commonly assumes paired or synchronized observations, yet independently collected datasets rarely provide such cor… | ↗Not briefed |
arxiv arxiv-ai | 25d | Can AI interview? Comparing AI-led to Human-led Chat Interviews | Large language model chatbots have been proposed as AI-interviewers for collecting qualitative interview data at scale, potentially br… | ↗Not briefed |
arxiv arxiv-ai | 25d | Grounded Credit Assignment in Reinforcement Learning for Multi-Turn Search Agents | Outcome-reward reinforcement learning such as GRPO gives every token of a search-agent rollout the same advantage, although the rollou… | ↗Not briefed |
arxiv arxiv-ai | 25d | Local proxies for memory-efficient training | Local learning algorithms can reshape resource requirements of artificial intelligence. The memory bottleneck of standard end-to-end b… | ↗Not briefed |
arxiv arxiv-ai | 25d | LexPilot: Benchmarking the Proactive Interaction Ability of Legal LLMs with Laypeople in Chinese Civil Case Consultation🔧 Elicit | Large Language Models (LLMs) have greatly assisted legal professionals. However, our empirical studies reveal that, when faced with un… | ↗Not briefed |
arxiv arxiv-ai | 25d | MetaCurate: Evolving Curation Filters for High-Quality Agentic Coding Data | Supervised fine-tuning on agent trajectories is an effective way to improve coding agents, but selecting useful training examples requ… | ↗Not briefed |
arxiv arxiv-ai | 25d | Information-Geometric Blind Spots of Conformal Test Martingales | Conformal test martingales (CTMs) are e-processes for sequentially testing exchangeability via p-values produced by a conformity funct… | ↗Not briefed |