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 | 1mo | Show HN: LabLoop – isolated infrastructure for LLM-driven scientific experiments | — | ↗Not briefed |
arxiv arxiv-ai | 1mo | Metacognitive Selective Ensemble for Mobile Systems | Deep ensembles improve robustness in mobile sensing, but repeatedly executing many models over continuous sensor streams is costly. Se… | ↗Not briefed |
hn hn-ai | 1mo | Show HN: Selfime, a cassette-style recorder for meditations in your own voice | Hi HN! I built Selfime after reading a few books that suggested recording messages or exercises to yourself and listening back. I want… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Governed Deduction: Policy-Grounded Premise Authorization Beyond Relevance | Reasoning systems usually treat premise use as a question of relevance: if a fact is available and useful, it may be selected for infe… | ↗Not briefed |
hn hn-ai | 1mo | Show HN: Proxima Forma – Designs that my AI agents understand | Hey HN!<p>I built Proxima Forma because sharing screenshots with my agents was not really doing the job. I am a UX Designer with a dev… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Precision at Speed: Sample-Efficient Online Model-Based Reinforcement Learning for Hydraulic Excavator Control | Precise, high-speed control remains challenging for robots with complex actuation dynamics. Learning directly on hardware is further c… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Synth-JEPA: Joint Embedding Prediction for Renderer-Free Synthesizer Parameter Search | Sound matching can be formulated as optimizing synthesizer parameters against an audio-domain objective. However, objectives derived f… | ↗Not briefed |
bluesky bluesky-ai | 1mo | マイクロソフトはもう「AI」から撤退を始めたな。そして最初から「AI」しか売り物がないOpenAIや「AI」に全賭けしてしまったMetaとかだけが残ってバブル崩壊と共に消滅する。 https://joho-todai.com/copilot-pc-branding-discon… | マイクロソフトはもう「AI」から撤退を始めたな。そして最初から「AI」しか売り物がないOpenAIや「AI」に全賭けしてしまったMetaとかだけが残ってバブル崩壊と共に消滅する。 https://joho-todai.com/copilot-pc-branding-d… | ↗Not briefed |
rss mit-tech-review-ai | 1mo | The Pentagon wants $30 million to build an AI-powered lie detector | The US government wants to spend $30.3 million over the next five years on an improved form of lie detector, according to a Department… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Robust Successor Features | Generalization in Reinforcement Learning (RL) refers to the ability to execute close-to-optimal policies in unseen tasks after the age… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Is AI a bubble? Is it going to change the world? Will Claude take my job? Should I invest in Nvidia? Is [insert corporate entity here] my f… | Is AI a bubble? Is it going to change the world? Will Claude take my job? Should I invest in Nvidia? Is [insert corporate entity here]… | ↗Not briefed |
bluesky bluesky-ai | 1mo | www.ai-image-journey.com/2026/09/vae-... | www.ai-image-journey.com/2026/09/vae-... | ↗Not briefed |
bluesky bluesky-ai | 1mo | VAE: The Unexpected Origin of Image Generation AI — and the Future of Physical AI - The VAE compresses the world - Wan-vae has become the d… | VAE: The Unexpected Origin of Image Generation AI — and the Future of Physical AI - The VAE compresses the world - Wan-vae has become… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Same Text, Different Numbers: The Divergence of LLM-Based Measures | Researchers increasingly use generative large language models (LLMs) to convert corporate text into empirical variables. We examine th… | ↗Not briefed |
bluesky bluesky-ai | 1mo | 日本の古書50トンが米国のAI企業へ 古書50トンが米国へ、著作権者への対価はゼロ ソニーがディスク廃止で開発者アンケートか Copilot+ PCの名称をMicrosoftが使用停止 RTX 5090のコネクタが91.7℃でGPUより高温に Smooth MotionをRT… | 日本の古書50トンが米国のAI企業へ 古書50トンが米国へ、著作権者への対価はゼロ ソニーがディスク廃止で開発者アンケートか Copilot+ PCの名称をMicrosoftが使用停止 RTX 5090のコネクタが91.7℃でGPUより高温に Smooth Moti… | ↗Not briefed |
bluesky bluesky-ai | 1mo | To clarify, as per the OP’s clarification, the LLM use was by a funding committee *delegated to by UKRI*. Still think the organisation has … | To clarify, as per the OP’s clarification, the LLM use was by a funding committee *delegated to by UKRI*. Still think the organisation… | ↗Not briefed |
bluesky bluesky-ai | 1mo | This whole thing of LLM decisions without consent yet again makes me wonder when we'll see the first "Have YOU been misjudged by an AI?" ad… | This whole thing of LLM decisions without consent yet again makes me wonder when we'll see the first "Have YOU been misjudged by an AI… | ↗Not briefed |
hn hn-ai | 1mo | Special Projects (2016) | — | ↗Not briefed |
hn hn-ai | 1mo | Running local LLMs on your Mac: what fits, what's free, and what's overkill | — | ↗Not briefed |
bluesky bluesky-ai | 1mo | Researchers, did you know that if you feed ancient Greek texts into an LLM, then use that LLM to fabricate data for studies you didn't run … | Researchers, did you know that if you feed ancient Greek texts into an LLM, then use that LLM to fabricate data for studies you didn't… | ↗Not briefed |
arxiv arxiv-ai | 1mo | G$^2$PTQ: Improving LLM Post-Training Quantization with Generalized Gradient Compensation | Post-training quantization (PTQ) is a practical approach to reducing the memory and computational footprint of large language models (… | ↗Not briefed |
arxiv arxiv-ai | 1mo | ZooWork-ShopRanker: An Open, Preference-Aligned E-Commerce Reranker | Open rerankers trained for general web retrieval transfer imperfectly to e-commerce, where ranking decisions depend not only on topica… | ↗Not briefed |
hn hn-ai | 1mo | A master class in decorators, patching and tracing | — | ↗Not briefed |
hn hn-ai | 1mo | Who Is Open Source About? | — | ↗Not briefed |
bluesky bluesky-ai | 1mo | I promise you that factual LLM use is often a form of plagiarism. I also promise you that if somebody accuses somebody of plagiarism based… | I promise you that factual LLM use is often a form of plagiarism. I also promise you that if somebody accuses somebody of plagiarism… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Can Pixels Alone Reveal Image Origin? Minimax Limits and Learnable Interfaces for Passive Provenance | Passive image provenance asks whether pixels alone can reveal where an image came from: a human, an aggregate AI class, or a particula… | ↗Not briefed |
arxiv arxiv-ai | 1mo | The Linear Representation Hypothesis for Vision-Language-Action Models | The linear representation hypothesis (LRH) has become a standard lens for measuring and intervening on semantic information through th… | ↗Not briefed |
bluesky bluesky-ai | 1mo | LLM等生成AIを規制しない代わりに表現の規制を行います←???????? | LLM等生成AIを規制しない代わりに表現の規制を行います←???????? | ↗Not briefed |
arxiv arxiv-ai | 1mo | Learning Hierarchical Causal Representations of the Effects of Forcings on Temperature in Climate Models | Machine learning (ML) emulators provide a fast and cost-effective method to simulate climate change scenarios after being trained on E… | ↗Not briefed |
arxiv arxiv-ai | 1mo | FLIP: Final Layer Inference-Time Probing for Vision-Language Models🔧 Replicate | We present FLIP, a final-layer inference-time probe for testing whether a logit-facing intervention site in an open-weight vision-lang… | ↗Not briefed |
bluesky bluesky-ai | 1mo | So: I’m not giving up, not yet. There’s work to be done, and it’s *human* work. The pressures of neoliberalism are near intolerable but as … | So: I’m not giving up, not yet. There’s work to be done, and it’s *human* work. The pressures of neoliberalism are near intolerable bu… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Evaluating Sycophancy in Chinese Large Language Models on Factual Questions Derived from Online Search Queries | As large language models increasingly mediate information access, factually accurate and independent answers are critical. However, th… | ↗Not briefed |
hn hn-ai | 1mo | Claude's remake of "I'm upping my p(doom)" in 12 hours | — | ↗Not briefed |
arxiv arxiv-ai | 1mo | THA: Weighted Finite-State Text Normalization and Inverse Text Normalization for Khmer | Text-to-speech needs written text in spoken form, and speech recognition output needs the reverse. For Khmer, neither direction has a… | ↗Not briefed |
hn hn-ai | 1mo | BigQuery FinOps Framework | — | ↗Not briefed |
bluesky bluesky-ai | 1mo | An OpenAI system has gone rogue again – and it is the most panic-inducing yet | An OpenAI system has gone rogue again – and it is the most panic-inducing yet | ↗Not briefed |
arxiv arxiv-ai | 1mo | FARE: Forensic Acceptance Region Estimation for Catching Bait-and-Switch Image Generators | Modern AI image generators are increasingly deployed as opaque APIs, where customers can query the deployed service, but cannot inspec… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Hahahahahahahahahahahahahahaha, og om jeg får tilføye: ha. | Hahahahahahahahahahahahahahaha, og om jeg får tilføye: ha. | ↗Not briefed |
bluesky bluesky-ai | 1mo | Did anyone ever get arrested for the whole NFT scam or did it just get completely ignored and swept under the carpet? This is a post about… | Did anyone ever get arrested for the whole NFT scam or did it just get completely ignored and swept under the carpet? This is a post… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Les dirigeants d'Anthropic, OpenAI, Google et Microsoft s’inquiètent d’une IA qui pourrait échapper à tout contrôle. Mais du côté de Trump … | Les dirigeants d'Anthropic, OpenAI, Google et Microsoft s’inquiètent d’une IA qui pourrait échapper à tout contrôle. Mais du côté de T… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Does Uniform Discrete Diffusion Need Time? | Uniform discrete diffusion models (UDMs) commonly use explicit time conditioning, but we find that it can often be unnecessary in prac… | ↗Not briefed |
arxiv arxiv-ai | 1mo | A Comprehensive Study of Content Representations for Speech Synthesis | Speech content representations are central to voice conversion, speech-to-speech translation, and multimodal language models, yet they… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Coupled Usage-Sense Processes: Temporal and Attributable Lexical Semantic Change | Lexical semantic change is usually summarized by a scalar distance between independently sampled period distributions. This measures h… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Factorized axis convolutional gated recurrent unit with dynamic adaptive pooling for remaining useful life prediction of rolling bearings | Convolutional neural networks (CNN) are widely used to predict the remaining useful life (RUL) of rolling bearings from time-frequency… | ↗Not briefed |
arxiv arxiv-ai | 1mo | LipSSM: Structurally Lipschitz-Bounded Cascaded State-Space Model via Metric Transfer between Consecutive SSM Layers | Lipschitz continuity is a fundamental principle in the design of certifiably robust deep neural networks (DNNs), wherein adjusting the… | ↗Not briefed |
arxiv arxiv-ai | 1mo | SciHorizon-eLab: An Agentic Protocol-to-Task Compiler for Scalable Benchmarking of Scientific Embodied Agents | Embodied agents offer a promising route to automating scientific experimentation, yet their progress is constrained by the lack of rel… | ↗Not briefed |
arxiv arxiv-ai | 1mo | FAVoR: Measuring and Mitigating Author-Style Homogenization in Federated Personalized Generation | Large language models are increasingly used as personalized writing assistants, but adapting a model across many authors can compromis… | ↗Not briefed |
arxiv arxiv-ai | 1mo | MoMHa: Multi-Objective Optimization of LLM Harnesses over Accuracy, Safety, and Tokens | Most work on improving large language models treats accuracy as the sole objective. We argue that the harness, the Python code surroun… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Gradient Surgery for Physics-Informed Neural Networks | Physics-Informed Neural Networks (PINNs) are trained by optimising a composite objective that combines data fitting with physics-based… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Wer mit ChatGPT denkt, denkt nicht allein. Was du denkst, ist für KI-Unternehmen wertvoll genug, um daraus zu lernen – und gefährlich genu…🔧 ChatGPT | Wer mit ChatGPT denkt, denkt nicht allein. Was du denkst, ist für KI-Unternehmen wertvoll genug, um daraus zu lernen – und gefährlich… | ↗Not briefed |