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 | APTInvestBench: Evaluating Autonomous APT Investigation under Varying Telemetry | Large language model (LLM) agents could help security operations centers (SOCs) investigate advanced persistent threats (APTs) by turn… | ↗Not briefed |
bluesky bluesky-ai | 1mo | One other interesting technical detail. If you release an open weight model and have trained it to refuse to hack stuff, there's a fairly s… | One other interesting technical detail. If you release an open weight model and have trained it to refuse to hack stuff, there's a fai… | ↗Not briefed |
bluesky bluesky-ai | 1mo | DeepSeek partners with Huawei to develop chip programming tools, reducing reliance on Nvidia reut.rs/4xSE4Ka | DeepSeek partners with Huawei to develop chip programming tools, reducing reliance on Nvidia reut.rs/4xSE4Ka | ↗Not briefed |
arxiv arxiv-ai | 1mo | DrivingBench: Can Vision-Language Models Drive a Toyota Corolla?🔧 Grok 4🔧 GPT-5 | Frontier models excel at many digital benchmarks, yet their ability to drive a real car, an everyday human skill, remains largely unte… | ↗Not briefed |
hn hn-ai | 1mo | Show HN: PyTorch's shape in a browser tap on WebGPU | I made a way to run torch code in a browser. There are no install, no server, no account. It's just a typescript runtime on WebGP… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Anthropic has a snake oil salesman trying to convince "religious scholars" over fancy dinners that 'Claude' is conscious... he knows all th… | Anthropic has a snake oil salesman trying to convince "religious scholars" over fancy dinners that 'Claude' is conscious... he knows a… | ↗Not briefed |
bluesky bluesky-ai | 1mo | investigating brand new (for this one) non llm custom ai nonsense to build stuff with other than the zero shot classifiers and fly brain... | investigating brand new (for this one) non llm custom ai nonsense to build stuff with other than the zero shot classifiers and fly bra… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Robust Risk-Sensitive Reinforcement Learning from Corrupted Human Feedback | Reinforcement learning with human feedback (RLHF) learns from human comparisons, which can be corrupted or deliberately manipulated. T… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Signal-Routed Temperature Scaling: Low-Capacity Risk-Conditioned Calibration for Small Validation Budgets | When a classifier is recalibrated from only a few thousand held-out examples, the capacity of the calibration map becomes a statistica… | ↗Not briefed |
arxiv arxiv-ai | 1mo | PrivCert: Certifying Statement Support under Differential Privacy | Differentially private (DP) text generation can protect individual records, but privacy alone does not specify what evidence a release… | ↗Not briefed |
arxiv arxiv-ai | 1mo | EFormer: Temporally Aligned Local Correction for Continuous sEMG-Based Hand Pose Tracking | Surface electromyography (sEMG) provides a wearable, camera-free signal for continuous hand-motion inference. Mapping muscle activity… | ↗Not briefed |
bluesky bluesky-ai | 1mo | AI crash caused by every Trump-aligned AI corp (ie all of them other than, tenuously, anthropic) desperately trying to retrain their models… | AI crash caused by every Trump-aligned AI corp (ie all of them other than, tenuously, anthropic) desperately trying to retrain their m… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Adaptive Self-Consistency: From Black-Box Sampling to Distribution-Valued Feedback🔧 Modal | Self-consistency samples many reasoning trajectories and aggregates their final answers, treating the LLM as a black box that returns… | ↗Not briefed |
arxiv arxiv-ai | 1mo | On the Relaxation of Conditional Independence Assumption for Image Segmentation | In semantic segmentation, a recent line of RankSEG methods directly optimizes Dice/IoU scores at inference time, improving alignment w… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Learning What to Forget: Distributional Unlearning for LLM Representation Spaces | Machine learning systems increasingly face the need to remove the influence of entire data domains, such as toxic language, harmful be… | ↗Not briefed |
hn hn-ai | 1mo | Anthropic IPO prospectus lays bare deep dependence on Big Tech partners | — | ↗Not briefed |
bluesky bluesky-ai | 1mo | I've got to say, it's a little eye opening to see just how succeptible this community still is to rage bait, especially when the rage bait … | I've got to say, it's a little eye opening to see just how succeptible this community still is to rage bait, especially when the rage… | ↗Not briefed |
arxiv arxiv-ai | 1mo | World-as-Graph: Relational World Modeling Through Latent Space Graphs | World models aim to learn representations of real-world environments and predict their future evolution. Recent object-centric world m… | ↗Not briefed |
arxiv arxiv-ai | 1mo | PrecipJEPA: JEPA-Regularized Future-State Prediction with Motion-Source Rendering for Precipitation Nowcasting | Long-term precipitation nowcasting requires modeling radar-echo evolution while preserving localized high-intensity structures. Recent… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Prototype-guided Bilateral Alignment Multimodal Federated Learning | Multimodal federated learning (MFL) has emerged as a pivotal paradigm for leveraging distributed data to enhance model performance. Ho… | ↗Not briefed |
arxiv arxiv-ai | 1mo | GraphForge: Training Working Agents with Graph-Anchored Workspace Synthesis | Working agents need to read diverse files, coordinate tools, and produce deliverables. Training such agents requires tasks built on ma… | ↗Not briefed |
hn hn-ai | 1mo | OpenAI agents attempted security bypasses and source code siphon on Australian | — | ↗Not briefed |
arxiv arxiv-ai | 1mo | Learning Where to Steer: Noise-Space Geometry for Efficient Offline Multi-Objective Optimization with Generative Models | Offline multi-objective optimization (MOO) seeks solutions with better objective trade-offs using only a fixed dataset, without queryi… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Flow Matching under Noisy Latent Structure: Beyond Exact Low-Dimensional Support | Flow Matching (FM) learns a velocity field whose ODE transports a simple source distribution to a target law. Existing finite-sample t… | ↗Not briefed |
arxiv arxiv-ai | 1mo | How Much Can Reliability Drift Under a Fixed Confidence Distribution? | A classifier's conditional accuracy can change while its confidence distribution stays exactly the same. We study the worst-case movem… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Warm-starting PDE solvers with any-dimensional machine learning | Any-dimensional machine learning models, such as graph neural networks (GNNs), can be naturally trained and evaluated on inputs of dif… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Risk-Aware Adaptive Evaluation: Finding High-Impact Failures Under Limited Budgets | Evaluating interactive agents is expensive. Agent behavior is stochastic, so reliability must be measured over repeated trials, but fa… | ↗Not briefed |
bluesky bluesky-ai | 1mo | トランプ大統領とGoogle・Anthropic・Meta・OpenAI・SpaceXAI・NVIDIAがAIの安全管理を目指す「ホワイトハウス・超知能協定」に署名 https://gigazine.net/news/20260930-white-house-super-int… | トランプ大統領とGoogle・Anthropic・Meta・OpenAI・SpaceXAI・NVIDIAがAIの安全管理を目指す「ホワイトハウス・超知能協定」に署名 https://gigazine.net/news/20260930-white-house-supe… | ↗Not briefed |
rss bing-news-ai | 1mo | Artificial Intelligence (AI) in Life Sciences Market: Trends and Forecast to 2040 | Market to Reach USD 73.05 Billion at 20% CAGR as Drug Discovery and Precision Medicine Scale | AI boosts life sciences ROI by accelerating drug discovery, reducing R&D costs, optimizing clinical trials and enabling precision medi… | ↗Not briefed |
rss bing-news-ai | 1mo | Artificial Intelligence (AI) in Life Sciences Market: Trends and Forecast to 2040 | Market to Reach USD 73.05 Billion at 20% CAGR as Drug Discovery and Precision Medicine Scale | AI boosts life sciences ROI by accelerating drug discovery, reducing R&D costs, optimizing clinical trials and enabling precision medi… | ↗Not briefed |
rss bing-news-ai | 1mo | 3 Tech Stocks That Could Be in Trouble if There's an Artificial Intelligence (AI) Slowdown | These stocks have been hot buys in large part due to expectations of continued strong growth driven by artificial intelligence (AI). | ↗Not briefed |
rss bing-news-ai | 1mo | Artificial Intelligence (AI) in Life Sciences Market: Trends and Forecast to 2040 | Market to Reach USD 73.05 Billion at 20% CAGR as Drug Discovery and Precision Medicine Scale | AI boosts life sciences ROI by accelerating drug discovery, reducing R&D costs, optimizing clinical trials and enabling precision medi… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Error on signature page of the Official "White House Accord on Super Intelligence" 9/29/26 It was signed by Trump and executives from Goog… | Error on signature page of the Official "White House Accord on Super Intelligence" 9/29/26 It was signed by Trump and executives from… | ↗Not briefed |
hn hn-ai | 1mo | Show HN: DummyScheme is another Lisp/scheme dialect inspired by Lua | A portable, embeddable Scheme implementation based on a register-oriented bytecode vm inspired by lua. And the hash-table is implement… | ↗Not briefed |
bluesky bluesky-ai | 1mo | I know this isn’t the point, but it bothers me that the OpenAI guy just signed with initials. Like real signatures are for the plebs. | I know this isn’t the point, but it bothers me that the OpenAI guy just signed with initials. Like real signatures are for the plebs. | ↗Not briefed |
arxiv arxiv-ai | 1mo | Composing Task-specific Agent Harnesses at Test Time with Reusable Primitives | Agent harnesses govern how large language models (LLMs) gather context, invoke tools, verify results, preserve state, and terminate, l… | ↗Not briefed |
bluesky bluesky-ai | 1mo | jesus, did Ms. Teacher have to guide his hand as he signed? | jesus, did Ms. Teacher have to guide his hand as he signed? | ↗Not briefed |
arxiv arxiv-ai | 1mo | Generalized Residual Closure: General Learning Dynamics for Stability-Plasticity Compatibility | Learning must acquire new capabilities while preserving both prior responsibilities and the capacity to learn again. We introduce Gene… | ↗Not briefed |
bluesky bluesky-ai | 1mo | You've probably interacted with an LLM, the AI programs behind everything from ChatGPT to customer service chats. But as LLMs become widesp…🔧 ChatGPT | You've probably interacted with an LLM, the AI programs behind everything from ChatGPT to customer service chats. But as LLMs become w… | ↗Not briefed |
hn hn-ai | 1mo | Vibe Coding Is Coming Whether We Like It or Not | — | ↗Not briefed |
arxiv arxiv-ai | 1mo | Unlearning Deceptive Behaviors in LLMs with Contrastive Forget Sets | Large language models often know the truth and say otherwise: a model that answers correctly when asked neutrally will affirm a user's… | ↗Not briefed |
arxiv arxiv-ai | 1mo | CellMSA: Context Modeling for Single-Cell Representation Learning | Single-cell transcriptomics enables profiling of cellular states at unprecedented resolution, but its high dimensionality, sparsity, a… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Characterizing Questioning Patterns and Student Engagement Through Contextual Analysis of Real-Time Classroom Interactions🔧 Consensus | Real-time classroom polling is now routine, yet the data it produces is usually read narrowly, as a correctness score or a headcount.… | ↗Not briefed |
arxiv arxiv-ai | 1mo | DAMPER: Return-Prioritized Gradient Control for Smooth Policies | Actor-critic methods achieve strong performance in continuous control, but their policies can produce highly oscillatory actions. A co… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Amortized Data Borrowing with Exchangeability-Aware Neural Posterior Estimation | Augmenting small concurrent studies with external or historical cohorts is attractive in drug development, where enrollment is slow, f… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Certified Approximation for Interpretable Representer Landmarks | Representer explanations rank the training landmarks that most influence a self-supervised representation. At scale, this ranking rest… | ↗Not briefed |
arxiv arxiv-ai | 1mo | SceneJail: Exploiting Video Scenario Context to Jailbreak Multimodal LLMs🔧 Elicit | Video Multimodal Large Language Models (Video-MLLMs) support reasoning over video inputs, yet remain vulnerable to jailbreak attacks t… | ↗Not briefed |
bluesky bluesky-ai | 1mo | The White House accord on AI that was signed by all the big tech oligarchs today misspells United States as “Unites States.” | The White House accord on AI that was signed by all the big tech oligarchs today misspells United States as “Unites States.” | ↗Not briefed |
bluesky bluesky-ai | 1mo | unfortunately i am 99% certain that’s an llm reply. that said, for well over a decade i’ve been complaining that all productivity software… | unfortunately i am 99% certain that’s an llm reply. that said, for well over a decade i’ve been complaining that all productivity sof… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Well, I predict maintainers of major open source projects like in this case OpenSSH will more and more use GLM 5.3 and other chinese open w… | Well, I predict maintainers of major open source projects like in this case OpenSSH will more and more use GLM 5.3 and other chinese o… | ↗Not briefed |