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 | 12d | Let the Fixer Decide: Training a Code Localizer from the Fixer's Likelihood | Code localizers are usually trained to predict the files a reference patch edits, but the code a fixer needs to read is not the same a… | ↗Not briefed |
hn hn-ai | 12d | Wattsec: Display macOS power usage in the menu bar | — | ↗Not briefed |
bluesky bluesky-ai | 12d | Llama 4 Maverick (17B-128E MoE) vs Gemini 2.5 Pro: Which LLM API actually protects your SaaS gross margin at 10,000 DAU? Here is the live …🔧 Llama 4 | Llama 4 Maverick (17B-128E MoE) vs Gemini 2.5 Pro: Which LLM API actually protects your SaaS gross margin at 10,000 DAU? Here is the… | ↗Not briefed |
arxiv arxiv-ai | 12d | Same Physics, Different Gradients: A State-of-the-Art Report on Differentiable Physics Simulation | Differentiable simulation propagates derivatives through a physics simulator, turning animation, system identification, control, and c… | ↗Not briefed |
bluesky bluesky-ai | 12d | When Jensen said he loves Chinese models, I thought it was about DeepSeek and Kimi and GLM, etc…… | When Jensen said he loves Chinese models, I thought it was about DeepSeek and Kimi and GLM, etc…… | ↗Not briefed |
hn hn-ai | 12d | Fun things to do with your VM/370 CE machine – Fun with big computers | — | ↗Not briefed |
arxiv arxiv-ai | 12d | GPU-Accelerated Non-Interactive Encrypted Neural Network Training with Batched Diagonal Packing | Training a neural network under homomorphic encryption requires transposes, batch reductions, optimizer updates, and repeated refreshe… | ↗Not briefed |
arxiv arxiv-ai | 12d | Beyond Flat Walks: Compositional Abstraction for Autoregressive Graph Generation | Autoregressive graph generators typically linearize graphs through flat node-edge walks, which can obscure higher-order structure in l… | ↗Not briefed |
bluesky bluesky-ai | 12d | "Sutton hatte bereits zuvor die Grenzen reiner Skalierung großer Sprachmodelle kritisiert, synthetische Daten als "großen Fehler" bezeichne… | "Sutton hatte bereits zuvor die Grenzen reiner Skalierung großer Sprachmodelle kritisiert, synthetische Daten als "großen Fehler" beze… | ↗Not briefed |
bluesky bluesky-ai | 12d | AIはサポート詐欺などの詐欺広告の文言を見抜けるか? 第2回目/編集室🖥 l.kooss.com/Qkb Gemini,ChatGPT,Copilot,Meta AI,Claude,DeepSeekなどのAIが人間以上に文言だけで詐欺広告を見抜ける水準になったか再びテストして…🔧 ChatGPT | AIはサポート詐欺などの詐欺広告の文言を見抜けるか? 第2回目/編集室🖥 l.kooss.com/Qkb Gemini,ChatGPT,Copilot,Meta AI,Claude,DeepSeekなどのAIが人間以上に文言だけで詐欺広告を見抜ける水準になったか再び… | ↗Not briefed |
bluesky bluesky-ai | 12d | La cosa es incluso peor! La app no está hecha con “código” está hecha con una plataforma no code que es simplemente “poner bloques” como s… | La cosa es incluso peor! La app no está hecha con “código” está hecha con una plataforma no code que es simplemente “poner bloques” c… | ↗Not briefed |
hn hn-ai | 12d | SHOW HN: GerGit - A Free Open-Source Git GUI Client for Linux Inspired by Fork | — | ↗Not briefed |
arxiv arxiv-ai | 12d | The balance between feature learning and collapse in dynamical systems-based generative models | Over the past decade, a growing body of work has established that implicit biases in learning dynamics fundamentally shape the solutio… | ↗Not briefed |
bluesky bluesky-ai | 12d | "Silicon Valley parecía haber tocado techo con la fantasía del metaverso(...) Pero el capitalismo tecnológico nunca muere: solo cambia de d… | "Silicon Valley parecía haber tocado techo con la fantasía del metaverso(...) Pero el capitalismo tecnológico nunca muere: solo cambia… | ↗Not briefed |
bluesky bluesky-ai | 12d | 「おしりの性能」とは🙄 gigazine.net/news/2026110... | 「おしりの性能」とは🙄 gigazine.net/news/2026110... | ↗Not briefed |
bluesky bluesky-ai | 12d | AIに「3DCGのおしり」を作らせて性能を比較するベンチマーク「LLM AssBench」が登場 https://gigazine.net/news/20261109-llm-assbench/ | AIに「3DCGのおしり」を作らせて性能を比較するベンチマーク「LLM AssBench」が登場 https://gigazine.net/news/20261109-llm-assbench/ | ↗Not briefed |
rss techmeme | 12d | Analysts warn Russia's Rassvet broadband satellite network could come online as early as 2027 after its troubled development, enabling more deadly drone strikes (Financial Times) | Financial Times: Analysts warn Russia's Rassvet broadband satellite network could come online as early as 2027 after its troubled deve… | ↗Not briefed |
hn hn-ai | 12d | Suitable for Taycan Turbo GT model carbon fiber tail | — | ↗Not briefed |
arxiv arxiv-ai | 12d | Do Time-Series Foundation Models Forecast Index-Futures Returns? A Preregistered Evaluation of Models, Metrics, and Input Representations | We ask whether the apparent forecasting skill of frozen time-series foundation models (TSFMs) on intraday index-futures returns is inc… | ↗Not briefed |
arxiv arxiv-ai | 12d | Visible Tails: Characterizing Adaptation to Unknown Moments in Heavy-Tailed Bandits | Heavy-tailed bandit learners without moment parameters face an obstruction: the rewards that carry an arm's mean can stay unseen. We a… | ↗Not briefed |
bluesky bluesky-ai | 12d | this is also a video about the impact of LLMs: github.com/hugelton/Fel... | this is also a video about the impact of LLMs: github.com/hugelton/Fel... | ↗Not briefed |
arxiv arxiv-ai | 12d | Good Views Are Not Enough: Global Counterexamples for Gaussian-Regularized JEPAs🔧 Modal | Good views do not suffice to guarantee scene invariance or uniqueness up to rotation at the global optima of Gaussian-regularized JEPA… | ↗Not briefed |
bluesky bluesky-ai | 12d | it's a claude subreddit. i'd be surpirsed if there were users of that subreddit who aren't using claude to write their posts. they sympathi… | it's a claude subreddit. i'd be surpirsed if there were users of that subreddit who aren't using claude to write their posts. they sym… | ↗Not briefed |
hn hn-ai | 12d | Ask HN: Do you still enjoy coding in C++? | In the recent years C++ received a lot of backlash due to its complexity and (perceived) feature creep. However, since C++11 I really… | ↗Not briefed |
rss techmeme | 12d | US federal prosecutors say businesses have lost $129M+ to scam calls using AI-generated voices to mimic company executives in the 18 months since July 2024 (Adeel Hassan/New York Times) | Adeel Hassan / New York Times: US federal prosecutors say businesses have lost $129M+ to scam calls using AI-generated voices to mimic… | ↗Not briefed |
arxiv arxiv-ai | 12d | One-Shot Federated Graph Learning via High-Fidelity Proxies and Transferability-Guided Collaboration | One-shot Federated Graph Learning (OFGL) has emerged as a communication-efficient paradigm for federated graph learning under strict b… | ↗Not briefed |
arxiv arxiv-ai | 12d | Toward Sustainable Agentic AI: Carbon Tradeoffs Across Iterative Self-Refinement and Tool-Augmented Coding🔧 GPT-5 | Agentic language model systems can improve performance either by iterating on their own outputs or by using external tools in grounded… | ↗Not briefed |
arxiv arxiv-ai | 12d | Lost in Rounding: bf16 Recurrent State Makes Hybrid Reasoning Models Think Longer | Hybrid reasoning models interleave attention with recurrent mixers such as Mamba-2 and Gated DeltaNet, whose state is cached between d… | ↗Not briefed |
arxiv arxiv-ai | 12d | CARE-PO: Controllable Tool-Call Efficiency for LLM Agents | External tool calls are a major running cost of an LLM agent, while its training objectives align almost exclusively with task complet… | ↗Not briefed |
bluesky bluesky-ai | 12d | Tech bros trying to sell me their latest AI LLM GPT bullshit. | Tech bros trying to sell me their latest AI LLM GPT bullshit. | ↗Not briefed |
arxiv arxiv-ai | 12d | System Optimizations for Cost-Efficient Agent Harnesses | AI agent harnesses play the critical role of providing LLMs with an interface to a computing environment. Recent studies show that AI… | ↗Not briefed |
bluesky bluesky-ai | 12d | i dont think i would enjoy having ddr4 ram sticks shoved down my throat but whatever you say anthropic | i dont think i would enjoy having ddr4 ram sticks shoved down my throat but whatever you say anthropic | ↗Not briefed |
rss techmeme | 12d | Sources: the EU sues IBM, French IT group Atos, and Italian contractor Leonardo over delays in developing its smart border system, which caused travel chaos (Laura Dubois/Financial Times) | Laura Dubois / Financial Times: Sources: the EU sues IBM, French IT group Atos, and Italian contractor Leonardo over delays in develop… | ↗Not briefed |
arxiv arxiv-ai | 12d | SCALE: Small-Model-First Concept Abstraction with LLM-Guided Experience Internalization | Autonomous agents must continually acquire reusable knowledge from experience without relying on costly large language model (LLM) rea… | ↗Not briefed |
bluesky bluesky-ai | 12d | respawnfirst.com/linux-script... | respawnfirst.com/linux-script... | ↗Not briefed |
arxiv arxiv-ai | 12d | Joi-Bench-Task: Benchmarking Task Coordination in Full-duplex Voice Agents through Interactive Simulation | Full-duplex voice agents enable natural, hands-free interaction, allowing users to refine requests, redirect ongoing work, and seek cl… | ↗Not briefed |
bluesky bluesky-ai | 12d | fact checked by real Claude patriots www.bbc.com/news/article... | fact checked by real Claude patriots www.bbc.com/news/article... | ↗Not briefed |
bluesky bluesky-ai | 12d | #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 | 12d | Forecast Value Add: Reward-Based Action Selection over Probabilistic Demand Forecasts | Strong machine-learning ensembles capture much of the trend, seasonality, and known events in demand, yet they can still understate pe… | ↗Not briefed |
arxiv arxiv-ai | 12d | OGDA: Ontology-Grounded Data Agents for Reliable Analysis over Heterogeneous Data Lakes | Full-pipeline data analysis over real-world data lakes requires LLM agents to locate and integrate information across large collection… | ↗Not briefed |
bluesky bluesky-ai | 12d | Tech bros trying to sell me their latest AI LLM GPT bullshit. | Tech bros trying to sell me their latest AI LLM GPT bullshit. | ↗Not briefed |
hn hn-ai | 12d | OpenIndiana 2026.10 Released with Updated Tools and Compilers, New Apps | — | ↗Not briefed |
bluesky bluesky-ai | 12d | Ich würde das zwar nicht überbewerten, aber das Geld sitzt inzwischen nicht mehr so locker. Ich bin skeptisch, was OpenAIs Position angeht.… | Ich würde das zwar nicht überbewerten, aber das Geld sitzt inzwischen nicht mehr so locker. Ich bin skeptisch, was OpenAIs Position an… | ↗Not briefed |
arxiv arxiv-ai | 12d | SkillDuet: Co-Evolving Models and Skills through Response Feedback | During reinforcement learning, experience can be both internalized in LLM parameters and retained externally as reusable textual skill… | ↗Not briefed |
hn hn-ai | 12d | Show HN: Glassbrain – Read your agent's brain, know what your agent knows | — | ↗Not briefed |
arxiv arxiv-ai | 12d | Perils and opportunities of AI for public climate services🔧 Consensus | The rapid integration of artificial intelligence (AI), particularly generative and machine-learning systems, in climate services devel… | ↗Not briefed |
rss techmeme | 12d | Chinese media denies Beijing's link to Jensen Huang's affair, accuses the US media of a fabricated spy narrative, and warns the issue may hurt trade relations (Bloomberg) | Bloomberg: Chinese media denies Beijing's link to Jensen Huang's affair, accuses the US media of a fabricated spy narrative, and warns… | ↗Not briefed |
arxiv arxiv-ai | 12d | FALC: An Artifact-Centric Life Cycle Model and Research Agenda for AI Forensics | AI systems increasingly appear in criminal investigations as tools, targets, and sources of evidence, yet AI forensics lacks a systema… | ↗Not briefed |
hn hn-ai | 12d | Show HN: Software Job Title Family Tree | Something I've always wanted to see, now made possible by AI: a family tree of computing / software job titles.<p>Hopefully… | ↗Not briefed |
arxiv arxiv-ai | 12d | Calibrating Document-Local Boundaries for Contextual Chunk Retrieval | Late Chunking encodes an entire document before pooling at chunk boundaries, preserving document context in local retrieval units. Chu… | ↗Not briefed |