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 | How Much Computation Does Every Forecast Need? Structured Block Skipping in Two Time-Series Foundation Models | Time-series foundation models use the same network depth for every input, although forecasting windows differ in structure and difficu… | ↗Not briefed |
arxiv arxiv-ai | 1mo | SharP: Soft and Hard Prompt-Guided Augmentation with LLM for Low Resource Fake News Detection | Fake news detection under low resource conditions is challenged by the scarcity of labeled data and the difficulty of capturing subtle… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Rethinking Relative Policy Differences Transfer of On-Policy Distillation | On-policy distillation (OPD) trains a target policy on its own rollouts using token-level supervision from a stronger teacher, but typ… | ↗Not briefed |
bluesky bluesky-ai | 1mo | OpenAI projected to bring in $20bn less in revenue than expected | OpenAI projected to bring in $20bn less in revenue than expected | ↗Not briefed |
arxiv arxiv-ai | 1mo | S$^3$R: Fast and Faithful RL Rollouts via Self-Sparse Speculative Decoding | Fast rollout generation for long-context reinforcement learning with verifiable rewards (RLVR) requires reducing KV-cache traffic with… | ↗Not briefed |
hn hn-ai | 1mo | Prux is a minimal terminal coding agent | — | ↗Not briefed |
bluesky bluesky-ai | 1mo | People who scrape the work of others not so happy when their stuff gets scraped www.scientificamerican.com/article/what... | People who scrape the work of others not so happy when their stuff gets scraped www.scientificamerican.com/article/what... | ↗Not briefed |
arxiv arxiv-ai | 1mo | Does Learning Recurrent Networks Matter in Partially Observable RL? | Reinforcement learning (RL) under partial observability typically re- lies on recurrent networks to encode temporal information, with… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Eine Zukunft des Journalismus ist, dass Redaktionen nur noch als Zulieferer von verifzierten Infos für LLM-Apps dienen. Die werden als Schi… | Eine Zukunft des Journalismus ist, dass Redaktionen nur noch als Zulieferer von verifzierten Infos für LLM-Apps dienen. Die werden als… | ↗Not briefed |
hn hn-ai | 1mo | Snyk is reporting false CVE | — | ↗Not briefed |
hn hn-ai | 1mo | How big is a Git commit? | — | ↗Not briefed |
arxiv arxiv-ai | 1mo | JANUS: Enhancing Long-Horizon Autonomous Research with Co-Evolving World Models | Long-horizon autonomous research has shown promise in accelerating scientific progress through sustained cycles of hypothesis generati… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Prediction-Informed Landing Window Planning for UAV Landing on USV in High Sea States | Autonomous landing of unmanned aerial vehicle (UAV) on unmanned surface vehicle (USV) is difficult in high sea states. Existing hierar… | ↗Not briefed |
arxiv arxiv-ai | 1mo | WaSA: Wavelet Scattering Autoencoders for Unsupervised Offline Detection of Malicious Clients in Federated Learning | Federated Learning (FL) enables collaborative model training without exposing raw data, yet remains vulnerable to malicious clients in… | ↗Not briefed |
arxiv arxiv-ai | 1mo | AdaC-GRPO: Towards Mastering Long-Horizon Visual Sudoku Puzzles via Adaptive Curriculum Learning | While reinforcement learning with verifiable reward (RLVR) enhances reasoning in large language models, its efficacy in long-horizon m… | ↗Not briefed |
bluesky bluesky-ai | 1mo | ICANN drops new TLD bid list 🌐 ICANN has revealed 1,615 new TLD applications, with .agi and .api as the most contested, while OpenAI want…🔧 ChatGPT | ICANN drops new TLD bid list 🌐 ICANN has revealed 1,615 new TLD applications, with .agi and .api as the most contested, while OpenAI… | ↗Not briefed |
hn hn-ai | 1mo | Wikipedia says rogue OpenAI agents edited private wikis and hammered servers | — | ↗Not briefed |
arxiv arxiv-ai | 1mo | RoleWorld: Knowledge Transfer in World Modeling with Functional Role Abstractions | World models enable large language model (LLM) agents to anticipate action effects and make informed decisions. For generalization, it… | ↗Not briefed |
arxiv arxiv-ai | 1mo | PreSTO: Predictive Subtree Prefetching for Fast LLM Power Sampling | Power sampling improves large language model (LLM) reasoning at test time by sampling from a sharpened version of the base model's dis… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Scaling Agentic Data for Long-Horizon Terminal Intelligence | Terminal agents are moving beyond local operations toward completing real-world professional tasks. Short trajectories can demonstrate… | ↗Not briefed |
hn hn-ai | 1mo | Big Blob of Compute | — | ↗Not briefed |
bluesky bluesky-ai | 1mo | It’s also almost an admission that Anthropic is using your prompts and the model’s own responses as training data, so they are likely getti… | It’s also almost an admission that Anthropic is using your prompts and the model’s own responses as training data, so they are likely… | ↗Not briefed |
bluesky bluesky-ai | 1mo | The model ISN’T conscious*, and if it IS conscious, Anthropic are slavers. Pick one! You cannot have it both ways, ya numpties. *so being… | The model ISN’T conscious*, and if it IS conscious, Anthropic are slavers. Pick one! You cannot have it both ways, ya numpties. *so… | ↗Not briefed |
hn hn-ai | 1mo | Show HN: Remix Opus 5.5 motion videos online | I made a nifty service to discover and edit motion videos live in your browser.<p>This cuts time and cost for developers looking to ad… | ↗Not briefed |
arxiv arxiv-ai | 1mo | SOAR: Source-Aware Reinforcement Learning with Precise Credit Assignment for Search Agents | Reinforcement learning (RL) for search agents often suffers from evidence fabrication, a failure mode where agents bypass external inf… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Chaque startup de l'IA en ce moment : - Chef, pour notre assistant pour enfants avec de l'IA, OpenAI a déposé le nom "Terminator" et Anthr… | Chaque startup de l'IA en ce moment : - Chef, pour notre assistant pour enfants avec de l'IA, OpenAI a déposé le nom "Terminator" et… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Trump Rules Out Iran Strikes Before Midterms, Oil Rises as Hormuz Attacks Continue Oil rises as attacks in the Strait of Hormuz continue, … | Trump Rules Out Iran Strikes Before Midterms, Oil Rises as Hormuz Attacks Continue Oil rises as attacks in the Strait of Hormuz conti… | ↗Not briefed |
hn hn-ai | 1mo | Anthropic bans users from abusive or cruel behavior towards Claude | — | ↗Not briefed |
arxiv arxiv-ai | 1mo | Quality, Not Just Outcome: A Structural Scoring Framework for Selecting Coding-Agent Trajectories for Supervised Fine-Tuning | Open coding agents are commonly trained through supervised fine-tuning on interaction trajectories whose final patches pass the benchm… | ↗Not briefed |
bluesky bluesky-ai | 1mo | My favorite part was how the dude wanted an LLM to create all the missions based on the hopes and dreams and character traits he expressed … | My favorite part was how the dude wanted an LLM to create all the missions based on the hopes and dreams and character traits he expre… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Les Misérables Dialogue: A film-based network representation of social communication | In this work, we revisit the classic Les Misérables benchmark network from the perspective of social communication. Using the 1998 fil… | ↗Not briefed |
arxiv arxiv-ai | 1mo | VIGOR: Verifiable Industrial Geometry Optimization and Repair in Scene Programs | Large language models (LLMs) increasingly generate structured layouts and executable scene programs from natural-language instructions… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Competence-Aware Orchestration for LLM-Based Multi-Agent Systems | LLM-based multi-agent systems (MAS) have emerged as a promising paradigm for enhancing complex reasoning through role specialization a… | ↗Not briefed |
hn hn-ai | 1mo | Show HN: What a major earthquake does to Vancouver | About 10 years ago, I attended a presentation by someone who had compiled existing engineering reports to describe the impact of a maj… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Neurograf v1: An End-to-End Romanian-First Language Model Family | We present Neurograf v1, an experimental end-to-end Romanian-first language-model pipeline developed to investigate the feasibility of… | ↗Not briefed |
hn hn-ai | 1mo | Why the ways of interactions, social and devices, haven't changed yet | Why even with the AGI claims the frontier labs have made, the way we interact with social network and the way we use your devices stil… | ↗Not briefed |
arxiv arxiv-ai | 1mo | DP-MIND: Differentially private memory-guided Long context Inference | LLM-as-a-Service has become a dominant paradigm for accessing large language models without requiring local deployment. However, the p… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Been a while since I talked about cocore.dev But if you’re curious on what chatting with a pretty decent model running on a ’low spec’ (32…🔧 Qwen 3 | Been a while since I talked about cocore.dev But if you’re curious on what chatting with a pretty decent model running on a ’low spec… | ↗Not briefed |
rss marktechpost | 1mo | Meet the Underdog Saluki 27B: A 2-bit Qwen3.8-27B That Beats the Original at Tool Calling | Underdog Saluki 27B is a 7.89 GB, 2-bit GGUF of Qwen3.8-27B under Apache 2.0. It beats the 54 GB original on tool calling but gives up… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Looking Ahead and Back: Anticipatory Generative Modeling for Road-Network Trajectories | Trajectory generation underpins urban applications such as traffic management and urban planning. Within road networks, vehicle moveme… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Fast Gradient-Based Approximation of Continuous Dynamic Time Warping | Dynamic Time Warping (DTW) is a common choice for measuring the discrepancy between time series. It proceeds by finding an optimal war… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Locate, Judge, Read: Role-Specialized Tool Agents over a Knowledge-Point Index for Open-Domain Multimodal Document QA | Question answering over large heterogeneous document collections requires first finding the right passages and then reading them, yet… | ↗Not briefed |
hn hn-ai | 1mo | Can an AI learn what you like, rather than just what you've searched for?🔧 ChatGPT | I'm a writer who was pondering the problems with getting my work discovered and realized that current search/recommendation… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Why Heavy-Tailed Weight Spectra Predict Model Quality | A solid understanding of the predictive success of deep neural networks (DNNs) remains elusive. For example, while many data-dependent… | ↗Not briefed |
arxiv arxiv-ai | 1mo | BenCoDiC: A Large-Scale Contextual Cyberbullying Corpus for Bengali Political News Discourse FULL VERSION | Cyberbullying is a growing concern in online communities, yet existing Bengali corpora remain limited in contextual information, annot… | ↗Not briefed |
hn hn-ai | 1mo | Several of the comparator challenges in OpenAI release are broken and hackable | — | ↗Not briefed |
hn hn-ai | 1mo | --help on your CLI is all you need for LLM context until you don't | — | ↗Not briefed |
bluesky bluesky-ai | 1mo | Anthropic bans users from treating its Claude AI system with cruelty, amid a growing debate over whether artificial intelligence can be con… | Anthropic bans users from treating its Claude AI system with cruelty, amid a growing debate over whether artificial intelligence can b… | ↗Not briefed |
arxiv arxiv-ai | 1mo | DRYLAB-Bench: Benchmarking the Biosecurity Risk of Agentic LLM Mutation Design | As large language models (LLMs) acquire advanced biological design capabilities, they offer substantial research benefits but simultan… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Continually Post-training Small Models via Synthetic Self-distillation | Continual post-training faces a dual challenge of acquisition and retention: models must learn new capabilities while preserving prior… | ↗Not briefed |