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 |
|---|---|---|---|---|
hn hn-ai | 18d | FaderBuddy – Plug-and-play motorized faders for ESPHome and more | — | ↗Not briefed |
arxiv arxiv-ai | 18d | From Outcomes to Steps: Reliable Success-Aware Critic Learning for Long-Horizon Language Agents | Reinforcement learning for long-horizon language agents requires identifying whether intermediate actions make eventual task success m… | ↗Not briefed |
arxiv arxiv-ai | 18d | A Head Camera Is (Almost) All You Need: An Electronics-Free UMI Glove Labelled and Observed by One Moving Camera | Handheld demonstration devices decouple data collection from the robot, but conventional designs burden the hand with a wrist camera,… | ↗Not briefed |
hn hn-ai | 18d | Token-based API guard Auth Bypass | — | ↗Not briefed |
arxiv arxiv-ai | 18d | A Radiation-Hardened, Fault-Aware Quantized Neural Network That Self-Heals | Deep learning (DL) models are increasingly being deployed into radiation-challenging environments in spacecraft, particle accelerators… | ↗Not briefed |
hn hn-ai | 18d | Show HN: Death Plan Switch, a self-hosted dead man's switch for solo devs | Hi HN,<p>I'm Peter Bishop and I've been working in graphics and web tech for over 25yrs.<p>I've recently launched Death… | ↗Not briefed |
arxiv arxiv-ai | 18d | Local Allow, Global Deny: Why Per-Action Authorization Does Not Compose in LLM Agents | Autonomous LLM agents act through tools under per-action authorization: a local guard screens each tool call in isolation. We show tha… | ↗Not briefed |
arxiv arxiv-ai | 18d | Learning Inter-Graph Interactions Between Heterogeneous Substructures of Chemical Systems | Complex chemical systems containing heterogeneous substructures are common in real-world applications. Various physical phenomena of t… | ↗Not briefed |
arxiv arxiv-ai | 18d | ABench: Adversarily Benchmarking Large Language Models across Taxonomized Values | Ensuring the alignment of Large Language Models (LLMs) with diverse cultural values is critical for their safe global deployment. Howe… | ↗Not briefed |
arxiv arxiv-ai | 18d | Auditing the Product, Not the Recursion: The Evaluation Evidence Behind Recursive Self-Improvement Claims | Recursive self-improvement (RSI) promises AI systems, such as self-evolving LLM agents, that become better at producing further improv… | ↗Not briefed |
bluesky bluesky-ai | 18d | I am very curious if Anthropic is gonna answer Google's challenge of "Our model admits it doesn't know, and it's MUCH less likely to halluc… | I am very curious if Anthropic is gonna answer Google's challenge of "Our model admits it doesn't know, and it's MUCH less likely to h… | ↗Not briefed |
arxiv arxiv-ai | 18d | Ask to Learn: Query-Based Collaborative Reinforcement Learning | Offline reinforcement learning is fundamentally limited by the coverage of the available data, while collecting additional trajectorie… | ↗Not briefed |
arxiv arxiv-ai | 18d | From Encoding to Generation: Towards Empathetic Large Language Models | Empathetic interaction involves two distinguishable capacities: representing another person's emotional state (cognitive empathy) and… | ↗Not briefed |
hn hn-ai | 18d | Ask HN: What's your SSH terminal workflow like? | I recently transitioned into doing most of my work via SSH running on an old server I've had lying around my house. Currently usi… | ↗Not briefed |
hn hn-ai | 18d | Show HN: VM 9.0.0, a mail reader for Emacs | VM is a mail reader that runs inside GNU Emacs, a once popular mail client brought back to life as a community project. VM 9.0.0 is th… | ↗Not briefed |
arxiv arxiv-ai | 18d | Self-Decomposing Diffusion for Code Generation | Diffusion language models (DLMs) are emerging as a promising paradigm for code generation, offering the potential for fast, parallel d… | ↗Not briefed |
rss bing-news-ai | 18d | Abiologics Appoints Per Jr. Greisen as Chief Artificial Intelligence Officer | AI protein design pioneer joins to advance first and only platform for de novo generation of synthetic D-protein therapeutics with sup… | ↗Not briefed |
rss bing-news-ai | 18d | Abiologics Appoints Per Jr. Greisen as Chief Artificial Intelligence Officer | AI protein design pioneer joins to advance first and only platform for de novo generation of synthetic D-protein therapeutics with sup… | ↗Not briefed |
arxiv arxiv-ai | 18d | Meta Expert: Cross-Layer Expert Sharing for Storage- and Retrieval-Efficient MoE | As model parameters scale up, sparsely activated Mixture-of-Experts (MoE) approaches have become essential for model scaling. However,… | ↗Not briefed |
arxiv arxiv-ai | 18d | WORTH Benchmark: Evidence Discovery under Not-Fully-Cooperative Information -- A Two-Episode Pilot | Existing reasoning evaluations typically provide task-relevant information to models as available premises, rarely examining concurren… | ↗Not briefed |
bluesky bluesky-ai | 18d | Not the easiest pill to swallow but the safety concerns of vibe coded software probably outweigh the ethical concerns even if the LLM was … | Not the easiest pill to swallow but the safety concerns of vibe coded software probably outweigh the ethical concerns even if the LLM… | ↗Not briefed |
arxiv arxiv-ai | 18d | Benefit and Cost of Random-Token Replacement under Repeated Data: A Known-Entropy Decomposition and Evidence from Ten Real Corpora | When unique pretraining data is scarce, models see the same tokens many times and are thought to memorize. Random-token replacement, a… | ↗Not briefed |
bluesky bluesky-ai | 18d | 深刻なインシデントが発生しているのに、明らかな定義の間違っている「AI(人工知能)」という言葉を使い続けて、LLMの技術的な特徴や欠点をきちんと調べないまま議論している事への危機感を、我が国の行政や司法や企業は持って欲しい。 たぶん、ほとんどのインシデントは「単純に勉強不足」で… | 深刻なインシデントが発生しているのに、明らかな定義の間違っている「AI(人工知能)」という言葉を使い続けて、LLMの技術的な特徴や欠点をきちんと調べないまま議論している事への危機感を、我が国の行政や司法や企業は持って欲しい。 たぶん、ほとんどのインシデントは「単純に勉… | ↗Not briefed |
arxiv arxiv-ai | 18d | Beyond Experience: Knowledge-Guided Evolution for NPU Kernel Generation | LLM agents have become capable kernel engineers on mainstream GPU stacks, but they struggle on emerging accelerators such as Ascend NP… | ↗Not briefed |
arxiv arxiv-ai | 18d | Sensitivity-Orthogonal Residual Learning for Adaptive Sideslip and Cornering Stiffness Estimation | Online adaptation of the cornering stiffness keeps a model-based sideslip estimator valid as tires change, but competes with learned m… | ↗Not briefed |
arxiv arxiv-ai | 18d | Bridging Digital Deception and Physical Fidelity: A Spatio-Temporal Attack on Cognitive Radars | As cognitive radars evolve towards the end-to-end processing of one-dimensional complex baseband sequences, deep networks face severe… | ↗Not briefed |
hn hn-ai | 18d | Show HN: Rouler – Steer Windows traffic per app or IP across multiple interfaces | Hi HN,<p>Rouler is "rules for your routes". It allows you to steer network traffic across multiple network interfaces on Win… | ↗Not briefed |
arxiv arxiv-ai | 18d | Beyond Model Weights: Sampling-Dependent Fairness in Diffusion Language Models | Diffusion language models (DLMs) are emerging as a practical alternative to autoregressive language models (ARLMs), yet their fairness… | ↗Not briefed |
arxiv arxiv-ai | 18d | TeCoP: Accelerating Diffusion Large Language Model Decoding via Tentative Context Pipelining | Diffusion large language models (dLLMs) predict multiple token positions per forward pass, yet block-wise decoders serialize generatio… | ↗Not briefed |
arxiv arxiv-ai | 18d | DuplexLedger: Benchmarking Transactional Memory in Full-Duplex Speech Agents | Full-duplex speech agents must update persistent memory while listening and speaking, yet it remains unclear which events should be st… | ↗Not briefed |
hn hn-ai | 18d | Show HN: An Open Source Aircraft | Good Luck, Have fun, Don't die. | ↗Not briefed |
arxiv arxiv-ai | 18d | Factual Access States in Diffusion Language Models: From Mechanism to Knowledge Editing and Retrieval | Factual recall is a lookup: some representation inside the network serves as the key, and what that key is made of decides which queri… | ↗Not briefed |
rss techmeme | 18d | An Australian minister says the eSafety Commissioner plans to sue Meta, TikTok, Snapchat, and YouTube over alleged breaches of the under-16 social media ban (Angus Whitley/Bloomberg) | Angus Whitley / Bloomberg: An Australian minister says the eSafety Commissioner plans to sue Meta, TikTok, Snapchat, and YouTube over… | ↗Not briefed |
bluesky bluesky-ai | 18d | Sometimes I get follows by AI people which is strange in and of itself, but sometimes they'll have extra wack shit in their bio like "semi-… | Sometimes I get follows by AI people which is strange in and of itself, but sometimes they'll have extra wack shit in their bio like "… | ↗Not briefed |
hn hn-ai | 18d | Show HN: Don't Get Nerf'd by Document Processing | We built doc extraction that owns the Pareto curve for almost 2 orders of magnitude.<p>Cost efficient structured outputs: - Near front… | ↗Not briefed |
arxiv arxiv-ai | 18d | Why Static Probes Fail: Context Determines Deception Feature Composition | With increasing autonomy, Large Language Models (LLMs) may engage in deceptive behaviors, such as scheming to pursue misaligned goals… | ↗Not briefed |
arxiv arxiv-ai | 18d | From POMDP Theory to Deep RL with Particle Filters | Model-based reinforcement learning in partially observable environments requires inferring hidden states, learning latent dynamics, an… | ↗Not briefed |
arxiv arxiv-ai | 18d | The Show-Up Attack: When Subagents Win Over Their Supervisor🔧 GPT-5 | Multi-agent systems are increasingly deployed to coordinate multiple large language model (LLM) agents. However, this also creates new… | ↗Not briefed |
arxiv arxiv-ai | 18d | When LLMs Manipulate: Measuring Behavioral, Social, and Temporal Effects | Large language models (LLMs) can influence user trust and decision-making through subtle linguistic cues, yet the behavioral effects o… | ↗Not briefed |
arxiv arxiv-ai | 18d | RaspAssistant as an Agentic Framework for Automating Adsorption Simulation Research in Porous Materials | Porous-material adsorption studies require coordinated structure preparation, force-field selection, simulation, validation, and inter… | ↗Not briefed |
arxiv arxiv-ai | 18d | Bayesian Neural Networks via Dirichlet Process Priors on the Data-Generating Distribution for Randomized Value Functions in Deep Reinforcement Learning | Bayesian neural networks (BNNs) offer a principled framework for quantifying epistemic uncertainty. However, standard weight-space pri… | ↗Not briefed |
arxiv arxiv-ai | 18d | Two Is Enough: Binary Classification Priors Suffice for Multi-Class In-Context Tabular Prediction | Tabular foundation models (TFMs) pretrain a transformer on large-scale synthetic priors so that a single forward pass yields zero-shot… | ↗Not briefed |
bluesky bluesky-ai | 18d | Soupçons d’espionnage chez OpenAI : un ingénieur saoudien maintenu en prison aux Etats-Unis | Soupçons d’espionnage chez OpenAI : un ingénieur saoudien maintenu en prison aux Etats-Unis | ↗Not briefed |
rss marktechpost | 18d | Inception Releases Mercury 3 Diffusion LLM at 1,068 Tokens per Second | Inception released Mercury 3, a diffusion LLM it says runs at 1,068 tokens per second with a 1M-token context. On Inception’s own tabl… | ↗Not briefed |
bluesky bluesky-ai | 18d | So glad all of TrumpPutin's diesel oil & data centers are going up in flames in Ruzzia. | So glad all of TrumpPutin's diesel oil & data centers are going up in flames in Ruzzia. | ↗Not briefed |
arxiv arxiv-ai | 18d | TRCM: Topology-Aware Route-Consistency Modeling for Robust Neural Vehicle Routing | Neural combinatorial optimization has enabled fast constructive solvers for the Capacitated Vehicle Routing Problem (CVRP), yet most a… | ↗Not briefed |
arxiv arxiv-ai | 18d | AudiPsy: Evidence-Grounded Multimodal Reasoning over Spoken Counseling Conversations for Depression Assessment | Depression assessment from counseling speech requires integrating textual and acoustic psychological evidence across turns, yet existi… | ↗Not briefed |
bluesky bluesky-ai | 18d | みなさま、こんにちは✨ お部屋でのんびりすごす日があってもいいよね🎶 【 参加者募集中 】 ブログ・日記・イラスト・写真・小説をBlueskyアカウントで投稿・販売できる複合型コンテンツサービス ➡️ atelier.fan #ブログ #日記 #イラスト #写真 #小説… | みなさま、こんにちは✨ お部屋でのんびりすごす日があってもいいよね🎶 【 参加者募集中 】 ブログ・日記・イラスト・写真・小説をBlueskyアカウントで投稿・販売できる複合型コンテンツサービス ➡️ atelier.fan #ブログ #日記 #イラスト #写… | ↗Not briefed |
bluesky bluesky-ai | 18d | Two Tech Bro capitalists offer up a "solution" to their AI making Americans unemployed in the near future?! fElon giving all children Space… | Two Tech Bro capitalists offer up a "solution" to their AI making Americans unemployed in the near future?! fElon giving all children… | ↗Not briefed |
arxiv arxiv-ai | 18d | Learned Readout Tokens for Extracting Large Language Model Confidence | Large language models (LLMs) are increasingly used within everyday applications where their responses have significant downstream impa… | ↗Not briefed |