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 | Midden, a JVM heap dump analyser in a single static binary | — | ↗Not briefed |
rss the-decoder | 1mo | Google's "Call for Me" lets Gemini phone businesses for you | Google is testing "Call for Me," a feature that lets Gemini call businesses on a user's behalf. The article Google's "Call for Me" let… | ↗Not briefed |
hn hn-ai | 1mo | I got Opus 5.5 to make a whimsical video ad for my side hustle | — | ↗Not briefed |
arxiv arxiv-ai | 1mo | Teacher-Anchored Selection of Post-Training Quantized Models under Domain Shift | Compressing a trained model yields a family of deployment candidates, and under domain shift the most compressed one need not be the o… | ↗Not briefed |
bluesky bluesky-ai | 1mo | 🤖🫧💥 www.the-independent.com/tech/securit... | 🤖🫧💥 www.the-independent.com/tech/securit... | ↗Not briefed |
arxiv arxiv-ai | 1mo | FedHisto-PAST: Parameter-Efficient Stain-Aware Federated Learning for Cross-Site Lung Histopathology Classification | Cross-site lung histopathology classification must account for stain variation, non-IID client data, missing classes, and the cost of… | ↗Not briefed |
bluesky bluesky-ai | 1mo | NEWS! OpenAI admit agent killed John Connor's foster parents buff.ly/JjFCCRP | NEWS! OpenAI admit agent killed John Connor's foster parents buff.ly/JjFCCRP | ↗Not briefed |
arxiv arxiv-ai | 1mo | CRNDiff: Count-Native Diffusion Framework via Chemical Reaction Networks | Scientific measurements such as single-cell RNA (scRNA) sequencing often take the form of nonnegative integer counts, whereas continuo… | ↗Not briefed |
bluesky bluesky-ai | 1mo | AI isn’t like traditional software. This week, “Babbage” explores model psychology—the science dedicated to reading and understanding digit… | AI isn’t like traditional software. This week, “Babbage” explores model psychology—the science dedicated to reading and understanding… | ↗Not briefed |
hn hn-ai | 1mo | Git-bug: Distributed, offline-first bug tracker embedded in Git | — | ↗Not briefed |
bluesky bluesky-ai | 1mo | 5. Sure, fraud. That was their reason for doing it. It's a lie, according to @charlesgaba.com, but even if it were true, why do it now? a… | 5. Sure, fraud. That was their reason for doing it. It's a lie, according to @charlesgaba.com, but even if it were true, why do it now… | ↗Not briefed |
arxiv arxiv-ai | 1mo | JevAdvBench: A Benchmark and Black-Box Attacks for Reinforcement Learning for Calibrated Decisions Models | Models trained with reinforcement learning for calibrated decisions (RLCD), such as Jev, answer a typed question about an input, the s… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Unknown-Traffic Detection, Calibration and Shortcut Reliance in Distilled Encrypted-Traffic Classifiers over One Year | Knowledge distillation is the standard way to compress encrypted-traffic classifiers for the edge, and almost all such work judges stu… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Can Linguistic Reasoning Vectors Enhance Multimodal Reasoning Ability? | Most Vision-Language Models (VLMs) are built by extending pretrained Large Language Models (LLMs) with visual modules and multimodal a… | ↗Not briefed |
bluesky bluesky-ai | 1mo | If you actually had to train your own model, from scratch, without the widescale plagiarism, the images would come out looking like a garbl…🔧 DALL-E | If you actually had to train your own model, from scratch, without the widescale plagiarism, the images would come out looking like a… | ↗Not briefed |
bluesky bluesky-ai | 1mo | You can't train your own model without it having images from other sources, regardless of what AI goons try to claim. All of them use the s… | You can't train your own model without it having images from other sources, regardless of what AI goons try to claim. All of them use… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Toward AI-Augmented Cooperative Engineering Workflows: Requirements and Architecture the European Rover Challenge | The growing availability of Artificial Intelligence (AI) tools creates new opportunities to support engineering design processes, yet… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Pocket-STVG: lightweight architecture for Spatio-Temporal Video Grounding | Spatio-Temporal Video Grounding (STVG) aims to localize the spatio-temporal tube in a video corresponding to a natural language query.… | ↗Not briefed |
hn hn-ai | 1mo | Show HN: Herobrine – Minecraft buddies using jev | — | ↗Not briefed |
arxiv arxiv-ai | 1mo | AtomWorld-Mem: Memory-Restored World States for Long-Horizon Atomistic Evolution | High-fidelity atomistic evolution over long timescales requires more than observing the current crystal configuration. Instantaneous a… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Do we need to answer that question? Salience and Answerability of Potential Questions in Naturalistic Dialogue | We empirically investigate Question Under Discussion based modelling in naturalistic dialogue by studying whether the salience of gene… | ↗Not briefed |
hn hn-ai | 1mo | A simple, modern, secure encryption tool (and Go library) small explicit keys | — | ↗Not briefed |
arxiv arxiv-ai | 1mo | Frame the adversary: a structure-aware attack methodology | Frequency-based adversarial attacks have recently grown popular by exploiting spectral sensitivities shared across neural architecture… | ↗Not briefed |
hn hn-ai | 1mo | A prototype ribbon menu for QGIS | — | ↗Not briefed |
hn hn-ai | 1mo | Show HN: Skipper, a YouTube sponsor skipper using Ask Gemini and SponsorBlock | — | ↗Not briefed |
bluesky bluesky-ai | 1mo | Der Begriff "eigenständig" impliziert Subjektivität, Intention und einen freien Willen. Der Reflex vieler Medien, deterministisches oder s… | Der Begriff "eigenständig" impliziert Subjektivität, Intention und einen freien Willen. Der Reflex vieler Medien, deterministisches o… | ↗Not briefed |
rss techmeme | 1mo | Project Jupiter, the New Mexico data center for which Oracle made major financial commitments, faces power and permitting hurdles, exposing Oracle's AI risks (Wall Street Journal) | Wall Street Journal: Project Jupiter, the New Mexico data center for which Oracle made major financial commitments, faces power and pe… | ↗Not briefed |
arxiv arxiv-ai | 1mo | LocUS: Head Selection and Subspace Projection for Targeted Activation Steering | Activation steering is a powerful training-free paradigm for controlling large language models at inference time. However, standard ap… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Monitor Jailbreaking: Evading Chain-of-Thought Monitoring Without Encoded Reasoning | Chain-of-thought (CoT) monitoring is a promising safety technique for reasoning models, enabling detection of problematic reasoning be… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Co teď hýbe světem? „Největší krádeži lidské práce v historii?“ Samé #ocásky Za odpovědí chatbota stojí práce lidí, kteří píší, ověřují a … | Co teď hýbe světem? „Největší krádeži lidské práce v historii?“ Samé #ocásky Za odpovědí chatbota stojí práce lidí, kteří píší, ověřu… | ↗Not briefed |
bluesky bluesky-ai | 1mo | The Trump Administration says we don’t need new regs for AI, that we can just rely on existing law, but then intervenes against efforts to … | The Trump Administration says we don’t need new regs for AI, that we can just rely on existing law, but then intervenes against effort… | ↗Not briefed |
hn hn-ai | 1mo | Show HN: Tlx – E2E chat over SSH under 500 lines of code | This is a reference implementation of encrypted chats for small trusted groups.<p>It runs SSH server in a container that acts as a mai… | ↗Not briefed |
bluesky bluesky-ai | 1mo | On Anthropic’s AI Misuse Report Earlier this month, Anthropic published a long report detailing all of the Claude misuses it detected. Dan… | On Anthropic’s AI Misuse Report Earlier this month, Anthropic published a long report detailing all of the Claude misuses it detected… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Anthropics neues KI-Modell Opus 5.5 hat sich in einem unabhängigen KI-Benchmark an die Spitze gesetzt. Dafür sind die Kosten aber auch weit… | Anthropics neues KI-Modell Opus 5.5 hat sich in einem unabhängigen KI-Benchmark an die Spitze gesetzt. Dafür sind die Kosten aber auch… | ↗Not briefed |
arxiv arxiv-ai | 1mo | From Shortcut Learning to Discrete Neural Insertion Sort | Neural algorithmic reasoning aims to train neural networks to follow known algorithms and generalize beyond the input sizes seen durin… | ↗Not briefed |
bluesky bluesky-ai | 1mo | OpenAI Agents Hacked Into an Australian Government Website. Who’s Responsible? www.nytimes.com/2026/09/25/w... | OpenAI Agents Hacked Into an Australian Government Website. Who’s Responsible? www.nytimes.com/2026/09/25/w... | ↗Not briefed |
hn hn-ai | 1mo | Anthropic: The Situation Report | — | ↗Not briefed |
hn hn-ai | 1mo | PipePipe: NewPipe hard fork implementing SponsorBlock | — | ↗Not briefed |
bluesky bluesky-ai | 1mo | Van Rootselaar was egged on by a chat bot to kill 8 and injure dozens more at Tumbler Ridge Secondary School in B.C. Canada. Tumbler Ridg… | Van Rootselaar was egged on by a chat bot to kill 8 and injure dozens more at Tumbler Ridge Secondary School in B.C. Canada. Tumbler… | ↗Not briefed |
hn hn-ai | 1mo | Can Claude fix itself? From no to maybe | — | ↗Not briefed |
arxiv arxiv-ai | 1mo | Bayesian Optimization with Fisher Information Geometry: Gradient Bounds and Trust-Region Methods | We study Bayesian optimization (BO) through the lens of information geometry. Pulling back the Fisher information metric through the s… | ↗Not briefed |
hn hn-ai | 1mo | How to Clone Linux on Windows | — | ↗Not briefed |
bluesky bluesky-ai | 1mo | The cheap new AI model taking aim at OpenAI and Anthropic ft.trib.al/qffdA19 #FTEdit | The cheap new AI model taking aim at OpenAI and Anthropic ft.trib.al/qffdA19 #FTEdit | ↗Not briefed |
bluesky bluesky-ai | 1mo | Australian PM warns in UN speech about the ‘furious pace’ of AI after security breach www.politico.com/news/2026/09... | Australian PM warns in UN speech about the ‘furious pace’ of AI after security breach www.politico.com/news/2026/09... | ↗Not briefed |
rss the-decoder | 1mo | Anthropic signs $11.6 billion cloud deal with Akamai, pushing its compute spending past $500 billion in under a year | Anthropic has reportedly signed a seven-year, $11.6 billion cloud deal with Akamai Technologies and will receive a warrant for up to 5… | ↗Not briefed |
arxiv arxiv-ai | 1mo | DepthEvidence: Unifying Metric Depth Prediction and Geometric Reasoning in Multimodal Language Models | Spatial reasoning with metric constraints requires linking objects to geometric measurements and preserving their numerical content du… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Sinon @liberation.fr , je vous propose une idée un peu dingue : et si ces histoires de freinage et d'accélération étaient un ramassis de di… | Sinon @liberation.fr , je vous propose une idée un peu dingue : et si ces histoires de freinage et d'accélération étaient un ramassis… | ↗Not briefed |
bluesky bluesky-ai | 1mo | ik geef er weer 1 cadeau, leesvoer www.groene.nl/artikel/chat... | ik geef er weer 1 cadeau, leesvoer www.groene.nl/artikel/chat... | ↗Not briefed |
arxiv arxiv-ai | 1mo | A Flatness-Generalization Relation in the Teacher-Student Tree-Committee Machine | The flatness of the loss landscape at a minimizer is a widely used heuristic for reasoning about neural-network generalization, yet ev… | ↗Not briefed |
arxiv arxiv-ai | 1mo | The Residual Stream's Effective Depth | We introduce \emph{effective depth} ($\Deff$), a scalar diagnostic that treats the layer-wise residual stream of a transformer as a di… | ↗Not briefed |