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 | 1mo | SBNO : Schrödinger Bridge Neural Operator for the Forward and Inverse Problems under Missing Information | Partial differential equation (PDE) surrogate modeling often relies on either dense observations or access to governing equations/phys… | ↗Not briefed |
hn hn-ai | 1mo | Lsnet – a fast TUI home network scanner in Rust | — | ↗Not briefed |
arxiv arxiv-ai | 1mo | JEPA-kMMVAE: CSI Prediction for Time-Varying Wireless Channels | This paper considers channel state information (CSI) prediction using neural networks for time-varying multi-antenna channels from noi… | ↗Not briefed |
bluesky bluesky-ai | 1mo | even "literal" translation involves choices, because words can have several meanings and languages organise things differently. An ai model… | even "literal" translation involves choices, because words can have several meanings and languages organise things differently. An ai… | ↗Not briefed |
bluesky bluesky-ai | 1mo | The AI lab insisted the decision was about a ‘breach of trust,’ not raising safety issues. www.theverge.com/ai-artificia... | The AI lab insisted the decision was about a ‘breach of trust,’ not raising safety issues. www.theverge.com/ai-artificia... | ↗Not briefed |
arxiv arxiv-ai | 1mo | Sustainable Cloud-Fog-Edge Systems for Smart Buildings: Algorithmic Adaptation under Memory Constraints and Dynamic Orchestration Strategies | The escalating environmental footprint of the Information and Communication Technol- ogy (ICT) sectorwith global Data Center electric-… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Structure-Aware Numerical Stabilization for Half-Precision Proximal Policy Optimization | Local online training on resource-constrained platforms requires low computational and training-state overhead. We study the numerical… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Claude Haiku 5.5 is a huge jump 🚀 Anthropic released Claude Haiku 5.5, touted as their cheapest, fastest, and most capable small model, c… | Claude Haiku 5.5 is a huge jump 🚀 Anthropic released Claude Haiku 5.5, touted as their cheapest, fastest, and most capable small mod… | ↗Not briefed |
arxiv arxiv-ai | 1mo | MORAD: Reinforcement Learning with Multi-Objective Rewards for RNA Inverse Design | We introduce MORAD, an online reinforcement learning framework for backbone-conditioned RNA diffusion models that trains a single shar… | ↗Not briefed |
arxiv arxiv-ai | 1mo | DAG Structure Learning with Node-Varying Heavy-Tailed Disturbances | DAG structures can be learned from observations by incorporating acyclicity via smooth constraints, such as DAGMA's log-determinant fu… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Agreement Without Alignment: Label-Free Rubric Clarification Increases LLM-Judge Consensus but Not Agreement with Humans🔧 Consensus | LLM judges are increasingly deployed as panels, and disagreement among panel members is often read as a symptom of an unclear rubric.… | ↗Not briefed |
rss theverge-ai | 1mo | OpenAI doubles down on decision to fire three AI safety researchers | OpenAI is standing firm on its decision to fire three safety researchers after an investigation found they committed "a significant br… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Sonic-AMG: Streaming Human Motion Generation from Spatial Audio | Modeling human reactions to spatial audio requires understanding both what is heard and where it comes from. Prior work generates moti… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Decoding Register Tokens Preserves Global Scene Details | Pretrained vision transformers learn structured image representations spanning spatial patch tokens and non-spatial tokens. Recent ana… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Fired OpenAI researchers say they were let go for 'prioritising safety' | Fired OpenAI researchers say they were let go for 'prioritising safety' | ↗Not briefed |
rss techmeme | 1mo | Sources: Netflix is preparing for layoffs that will impact ~5% of employees, or ~850 jobs, amid pressure over weakened engagement and a depressed stock price (Matthew Belloni/Puck) | Matthew Belloni / Puck: Sources: Netflix is preparing for layoffs that will impact ~5% of employees, or ~850 jobs, amid pressure over… | ↗Not briefed |
arxiv arxiv-ai | 1mo | PRIME: Pre-Retrieval Index Modeling Enhancement for Structured Document Retrieval-Real | High-precision retrieval over external knowledge bases is essential for LLM agents to operate effectively. However, existing Informati… | ↗Not briefed |
hn hn-ai | 1mo | Once: Cache CLI Commands | — | ↗Not briefed |
arxiv arxiv-ai | 1mo | Recovering Irreducible Three-Body Interactions with Lagrangian Neural Networks | Lagrangian Neural Networks (LNNs) learn mechanical systems from trajectory data by parameterizing a scalar Lagrangian and deriving the… | ↗Not briefed |
bluesky bluesky-ai | 1mo | So many philosophy PhDs on staff an nobody told them that you can't be "cruel" to a bunch of statistics. Anthropic bans ‘abusive or cruel … | So many philosophy PhDs on staff an nobody told them that you can't be "cruel" to a bunch of statistics. Anthropic bans ‘abusive or c… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Do Prompt Optimizers Break Requirements They Are Not Scored On? A Preregistered Study🔧 Elicit | Prompt optimizers rewrite a prompt to raise one target metric, but prompts also state requirements the metric does not score, such as… | ↗Not briefed |
hn hn-ai | 1mo | AmberCell is a high-interaction honeypot: it runs full, real services | — | ↗Not briefed |
arxiv arxiv-ai | 1mo | BoundaryEvo: Data-Model Co-Evolution at the Capability Boundary for Long-Chain Reasoning Distillation | Distilling long-chain reasoning trajectories from large teacher models provides an effective way to improve the reasoning ability of s… | ↗Not briefed |
hn hn-ai | 1mo | Show HN: Long term Memory and 50M token window for LLM | — | ↗Not briefed |
rss bing-news-ai | 1mo | Trump Wants to Rename Artificial Intelligence | US President Donald Trump wants to replace the term "artificial intelligence" (AI) with "super intelligence" (SI), arguing that the wo… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Pretty interesting Harvard difference-in-differences study of LLM tool productivity gains in software engineering, based on Jellyfish (a th… | Pretty interesting Harvard difference-in-differences study of LLM tool productivity gains in software engineering, based on Jellyfish… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Writing is thinking. When you outsource your writing to an LLM, you rob yourself of the chance to think. Thomas Ptacek advice is: use AI a… | Writing is thinking. When you outsource your writing to an LLM, you rob yourself of the chance to think. Thomas Ptacek advice is: use… | ↗Not briefed |
arxiv arxiv-ai | 1mo | MHR-Router: Multimodal Heterogeneous Retrieval-Enhanced Adaptive Routing for VLM Inference | Vision-language models (VLMs) exhibit complementary capabilities, yet selecting the most suitable expert for a given image-question qu… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Beyond the Final Embedding: Retrieving Memories from Token-Level Evidence | Memory-augmented assistants retrieve evidence from stored conversations and documents to answer queries. Most embedding-based retrieve… | ↗Not briefed |
arxiv arxiv-ai | 1mo | CopRoute: Dependence-Aware Conformal Routing for Language Models | Set-based language model routing aims to increase the coverage of the event where at least one model answers a query correctly while c… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Agent4Spec: An LLM-based Agentic System to Optimize API Specifications through Automated Validation and Market Analysis | API specifications serve as critical documentation for modern software development, directly impacting integration efficiency and clie… | ↗Not briefed |
hn hn-ai | 1mo | OpenAI defends decision to fire researchers | — | ↗Not briefed |
rss the-decoder | 1mo | Anthropic's Claude Science creates the first complete ultraviolet map of the sky | Astrophysicist Brice Ménard of Johns Hopkins University used Anthropic's Claude Science to map the entire sky in ultraviolet light for… | ↗Not briefed |
bluesky bluesky-ai | 1mo | 女を殴りがちなＡＩ真木Claude | 女を殴りがちなＡＩ真木Claude | ↗Not briefed |
hn hn-ai | 1mo | Tell HN: YouTube is punishing Incognito Mode users | YouTube has implemented hair trigger "anti-bot" measures that I suspect strictly limit the number of session cookies issued… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Forgetting is Not Always Bad: A Neuro-Inspired Memory Repair Mechanism for Poisoned LLM Agents | Large language model (LLM) agents are highly vulnerable to memory-poisoning attacks. Existing defenses primarily rely on external modu… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Flew too close to the sun? sifted.eu/articles/ex-... | Flew too close to the sun? sifted.eu/articles/ex-... | ↗Not briefed |
arxiv arxiv-ai | 1mo | CV-OPD: Contrastive Vision On-policy Distillation | On-policy distillation (OPD) reveals versatility and convenience during VLM post-training, providing dense supervision for student fro… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Data-Adaptive Mahalanobis Metric Learning for Cross-Head Attention in Transformers | In this paper, we propose Data-Adaptive Mahalanobis Cross-Head Attention (DAMCHA), a simple and effective attention mechanism that gen… | ↗Not briefed |
hn hn-ai | 1mo | TokenRouter: A serving engine for token-level LLM routing | — | ↗Not briefed |
arxiv arxiv-ai | 1mo | Neural Surrogate Modeling of Ultrasonic Wave Propagation in Heterogeneous Media | Neural surrogates provide computationally efficient auto-regressive modeling of wave propagation, enabling e.g. fast inverse solutions… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Nog nooit Claude gebruikt, maar ik download haast de app om hem uit te schelden. www.theguardian.com/technology/2... | Nog nooit Claude gebruikt, maar ik download haast de app om hem uit te schelden. www.theguardian.com/technology/2... | ↗Not briefed |
bluesky bluesky-ai | 1mo | If your rationale why you desperately want to join my seminar I didn‘t even ask for is also obviously written by ChatGPT it‘s not a great f…🔧 ChatGPT | If your rationale why you desperately want to join my seminar I didn‘t even ask for is also obviously written by ChatGPT it‘s not a gr… | ↗Not briefed |
arxiv arxiv-ai | 1mo | CAST: Credit Alignment for Cross-Tokenizer On-Policy Distillation | Cross-tokenizer on-policy distillation (OPD) requires aligning teacher and student sequences before their scores can be compared. Deco… | ↗Not briefed |
bluesky bluesky-ai | 1mo | 3 months later vibekiddies can just point Artex (open model pentesting toolkit) at Korean banks to pwn them www.crowdstrike.com/en-us/blog… | 3 months later vibekiddies can just point Artex (open model pentesting toolkit) at Korean banks to pwn them www.crowdstrike.com/en-us… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Tabular Foundation Models Are Effectively Shallow | In-context-learning Tabular Foundation Models (TFMs) are uniquely suited to scenarios with limited resources or privacy constraints, a… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Vision Transformers Learn Word Meaning from Images Alone | Can a vision model learn what words mean from pixels alone? We show that the two largest DINOv3 vision transformers organize words ren… | ↗Not briefed |
rss mit-tech-review-ai | 1mo | We’re putting too much faith in AI’s ability to say no | Ever since people first seriously contemplated giving machines an intelligence modeled on our own, there has never been any question t… | ↗Not briefed |
hn hn-ai | 1mo | Show HN: Stećak – a lightweight terminal inspired by Bosnian medieval tombstones | I've been working on Stećak, a lightweight terminal emulator inspired by Ghostty and Alacritty.<p>The idea came from stećci (sing… | ↗Not briefed |
arxiv arxiv-ai | 1mo | What Do Audio-Language Judges of Generated Music Listen For? Four Probes with Controls from Expert Ratings | Open audio-language models are used as judges of generated music: given a clip and its text prompt, the model rates the clip, and the… | ↗Not briefed |