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 |
|---|---|---|---|---|
bluesky bluesky-ai | 1mo | 'ICE wil busjes ombouwen tot “mobiele detentiecentra” waarin arrestanten op straat kunnen worden vastgehouden. De plannen zeggen weinig ove… | 'ICE wil busjes ombouwen tot “mobiele detentiecentra” waarin arrestanten op straat kunnen worden vastgehouden. De plannen zeggen weini… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Supervise What Decides Success: Criterion-Aligned Auxiliary Losses for Latent World-Model Planning | Latent world models plan by scoring candidate action sequences with distances in latent space. However, task success is judged by phys… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Have an LLM Write Your Anomaly Detector: Autonomous Discovery of Compact, Interpretable Detectors for Time Series | Time-series anomaly detection trades off predictive accuracy, computational efficiency, and interpretability. We use a large language… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Reputation, Strategy, and Emotion Effects on Generative AI Cooperation: A Comparison Across Reasoning and Non-Reasoning Models🔧 GPT-5 | As generative AI (Gen AI) systems take on increasingly autonomous roles in economically and socially consequential interactions, under… | ↗Not briefed |
arxiv arxiv-ai | 1mo | AGO AI Quality Gate: Evidence-First Release Decisions for Retrieval-Augmented Generation | Enterprises adopting retrieval-augmented generation (RAG) face a recurring operational decision: promote, revise, or block a system ve… | ↗Not briefed |
bluesky bluesky-ai | 1mo | I think the most likely AI apocalypse looks a lot more like @aptshadow.bsky.social 's novel Service Model than Terminator. And that's terri… | I think the most likely AI apocalypse looks a lot more like @aptshadow.bsky.social 's novel Service Model than Terminator. And that's… | ↗Not briefed |
bluesky bluesky-ai | 1mo | The single most important person to my building a *evolving* mental model of generative AI is @colin-fraser.net. As these tools change so t… | The single most important person to my building a *evolving* mental model of generative AI is @colin-fraser.net. As these tools change… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Dependency-Aware Reward Shaping for Agentic Reinforcement Learning | When training large language models with reinforcement learning, terminal rewards provide little guidance about which steps matter. Co… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Autoregressive Drillhole Modelling Under Distribution Shift | Autoregressive modelling has achieved remarkable success in language and sequence tasks by learning to predict future states from prev… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Never fails to amaze how socially misanthropic the tech bros are. Instead of ensuring #AI models are constrained by universal norms, such a… | Never fails to amaze how socially misanthropic the tech bros are. Instead of ensuring #AI models are constrained by universal norms, s… | ↗Not briefed |
hn hn-ai | 1mo | We'll limit access to old Reddit to users that haved used it in the last 6 month | — | ↗Not briefed |
bluesky bluesky-ai | 1mo | *Consider This" podcast from NPR: Interview with Jacob Coxon, resigning from Anthropic spurred by perceived dangers of AI. Triggered my qu… | *Consider This" podcast from NPR: Interview with Jacob Coxon, resigning from Anthropic spurred by perceived dangers of AI. Triggered… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Low-Budget Active Learning through Entropic Optimal Transport | We consider low-budget active learning, which consists of selecting a limited number of points, the coreset, such that a model can be… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Yesterday my boss suggested that Claude might be able to help me with something I've been doing professionally for 25 years. I do it incred… | Yesterday my boss suggested that Claude might be able to help me with something I've been doing professionally for 25 years. I do it i… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Federated Agent Optimization | Large language model (LLM) agents increasingly operate in private environments and accumulate valuable experience from task execution,… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Counterfactual Generation via Flow Matching: Coupling-Sensitive End-to-End Rates | Counterfactual generation seeks to sample outcomes under a hypothetical intervention or decision using observational data collected un… | ↗Not briefed |
arxiv arxiv-ai | 1mo | When Does Exercise-Specific Joint Selection Help? An Audit of Evaluation and Control Design | Exercise-specific joint selection can improve skeleton-based correctness classification, but what does that gain establish? We audit 1… | ↗Not briefed |
hn hn-ai | 1mo | Tarve: Native UI Framework for Bun and TypeScript Powered by Rust | — | ↗Not briefed |
bluesky bluesky-ai | 1mo | A recent disclosure of Amazon’s internal “employee relations” documents reveals just how extreme Amazon’s methods of warehouse surveillance… | A recent disclosure of Amazon’s internal “employee relations” documents reveals just how extreme Amazon’s methods of warehouse surveil… | ↗Not briefed |
arxiv arxiv-ai | 1mo | ReCast: Contract-Preserving Protection for Fixed-Interface Multimodal Reasoning | Remote multimodal models offer strong numerical reasoning capabilities over charts and speech, but sending private inputs risks exposi… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Researchers found that even when users rejected non-essential cookies, Perplexity, DeepSeek, Gemini, Copilot, ChatGPT and Claude still conn…🔧 ChatGPT | Researchers found that even when users rejected non-essential cookies, Perplexity, DeepSeek, Gemini, Copilot, ChatGPT and Claude still… | ↗Not briefed |
rss marktechpost | 1mo | NVIDIA Releases Kumo Tabular: Open Tabular Foundation Models That Predict New Rows in a Single Forward Pass | NVIDIA has released Kumo Tabular, a new family of tabular foundation models (TFMs) for classification and regression. If you have foll… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Fully Online Decentralized Learning in Stochastic Games with Unknown Independent Chains | We consider stochastic games with independent controlled chains and unknown transition kernels, where players observe only their local… | ↗Not briefed |
hn hn-ai | 1mo | Binjgb: Gameboy emulator implemented in C, that also runs in the browser | — | ↗Not briefed |
bluesky bluesky-ai | 1mo | Cuidado con lo que compartes con la IA, especialmente si no es información tuya y es inédita, como manuscritos de tesis, en peer review, et… | Cuidado con lo que compartes con la IA, especialmente si no es información tuya y es inédita, como manuscritos de tesis, en peer revie… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Temporally-Resolved Token Attribution Reveals the Generation Dynamics of Diffusion Language Models | This work presents Diffusion Layer Integrated Gradients (DLIG), a token attribution method for diffusion language models (DLMs) that e… | ↗Not briefed |
bluesky bluesky-ai | 1mo | How I’d learn AI SEO if I starting today (2026) - youtu.be/TApKjObgMfo?... #seo #geo #llm #ai | How I’d learn AI SEO if I starting today (2026) - youtu.be/TApKjObgMfo?... #seo #geo #llm #ai | ↗Not briefed |
arxiv arxiv-ai | 1mo | Rethinking the Information Bottleneck: Structured Decomposition under Label-Induced Partitions | Standard information bottleneck (IB) regularization constrains representations via a single scalar I(Z;X), implicitlytreating all info… | ↗Not briefed |
hn hn-ai | 1mo | Let skills in Gemini tackle your most repetitive tasks | — | ↗Not briefed |
arxiv arxiv-ai | 1mo | CAGE-NAS: Certified Functional Descent for Efficient Model Growth | The progressive growth of neural networks requires deciding when the current representation remains sufficient for optimization and wh… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Learning Rate Transfer for Hybrid Transformer-SSM Architectures | We study learning rate (LR) scaling for hybrid architectures combining Transformer and State-Space Model (SSM) blocks, a class adopted… | ↗Not briefed |
arxiv arxiv-ai | 1mo | HeadEdit: Calibrating Language Model Behavior Through the Frozen Unembedding Matrix | Alignment does not eliminate behavioral errors in language models. Models may still refuse benign requests, call unnecessary tools, or… | ↗Not briefed |
rss latent-space | 1mo | [AINews] Gemini 4 Argon: GDM’s answer to Astra/Fable, with 1M output | ... but you can’t try it yet unless you are “government users and trusted cyber defenders in the Fairwind Program” | ↗Not briefed |
arxiv arxiv-ai | 1mo | Detect, Explain, Interpret: An End-to-End Benchmark for Time Series Anomaly Detection, Explainability and Interpretability | Time Series Anomaly Detection has received increasing attention, driven by the growing availability of complex time series data. This… | ↗Not briefed |
arxiv arxiv-ai | 1mo | CineMR: Tool-Integrated Vision-Language Reasoning for Quantitative Cardiac MRI Assessment | Cardiovascular magnetic resonance (CMR), including cine imaging, is a reference standard for the noninvasive assessment of cardiac mor… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Persistent Depth Ordering amid Shifting Block-Bypass Responses in Language Model Pretraining | Layer interventions are widely used to probe the internal organization of language models, yet most analyses examine a single training… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Of course I would never use an LLM as a calculator. But it’s an existence proof. The model can answer at least some questions it hasn’t mem… | Of course I would never use an LLM as a calculator. But it’s an existence proof. The model can answer at least some questions it hasn’… | ↗Not briefed |
hn hn-ai | 1mo | OpenAI's Brockman Backs Out of Pledged $25M Donation to AI Super Pac | — | ↗Not briefed |
bluesky bluesky-ai | 1mo | Google AI tops CDC flu forecast Google’s Google_SAI-FluEns model topped the CDC’s 2025-2026 FluSight evaluation for U.S. flu hospital admi… | Google AI tops CDC flu forecast Google’s Google_SAI-FluEns model topped the CDC’s 2025-2026 FluSight evaluation for U.S. flu hospital… | ↗Not briefed |
arxiv arxiv-ai | 1mo | My FAULT: Self-Diagnosis as Credit Assignment in Self-Evolving Agentic Reinforcement Learning | Agentic reinforcement learning (RL) has emerged as a powerful approach for training large language model agents on multi-step tasks, y… | ↗Not briefed |
hn hn-ai | 1mo | AWS just got a lot easier to use | — | ↗Not briefed |
bluesky bluesky-ai | 1mo | 120 days to mAp ThE fUtUrE oF wArFiGhTiNg with Musk, Luckey and Newt Gingrich. Maybe they can also feed all of Tom Clancy's books into an L… | 120 days to mAp ThE fUtUrE oF wArFiGhTiNg with Musk, Luckey and Newt Gingrich. Maybe they can also feed all of Tom Clancy's books into… | ↗Not briefed |
rss techmeme | 1mo | A profile of Xbox CEO Asha Sharma, who insists that "Xbox is not for sale" despite mass layoffs and divested studios since inheriting the flailing division (Zachary Small/New York Times) | Zachary Small / New York Times: A profile of Xbox CEO Asha Sharma, who insists that “Xbox is not for sale” despite mass layoffs and di… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Looping Beyond Twice: A Scalable Recipe for Looped Mixture-of-Experts | Looped Transformers introduce recurrent depth as a new scaling axis for LLMs: by repeatedly applying shared Transformer blocks, they i… | ↗Not briefed |
rss techmeme | 1mo | SemiAnalysis estimates ~90% of Anthropic's business comes from agentic AI, while sources say nearly 25% of its revenue in 2025 came from just two clients (Financial Times) | Financial Times: SemiAnalysis estimates ~90% of Anthropic's business comes from agentic AI, while sources say nearly 25% of its revenu… | ↗Not briefed |
arxiv arxiv-ai | 1mo | BanglaDial-Abuse: A Corpus-Grounded Dataset for Regional Dialect Identification in Abusive Bangla Text | Regional linguistic variation remains an important challenge for Bangla natural language processing, particularly in informal and non-… | ↗Not briefed |
rss simon-willison | 1mo | Quoting Matthew Green | — | ↗Not briefed |
arxiv arxiv-ai | 1mo | When Is Deletion Ordering Tractable? From Update Dynamics to Permutation Structure | Given a fixed set of pending deletion requests, retraining from scratch after each request is prohibitive, so a prescribed request-wis… | ↗Not briefed |
hn hn-ai | 1mo | Show HN: Animated UI components you can preview and install in one command | Scroll to see animations. Hover to copy the install command. Get editable source code. Menus, loaders, transitions, carousels, and mor… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Parameter-Efficient Distributionally Robust Adaptation of Tabular Foundation Models under Subpopulation Shift | Despite strong mean accuracy, tabular foundation models (TFMs) can perform poorly on underrepresented groups under subpopulation shift… | ↗Not briefed |