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 | Chaining Skills to Hijack LLM Agents🔧 GPT-5 | LLM agents use skills to improve performance on specialized tasks. To complete a user request, an agent may invoke several skills in s… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Trump calls AI safety fears a “hoax” while an Anthropic co-founder says companies may need a mandatory, independently verified kill switch.… | Trump calls AI safety fears a “hoax” while an Anthropic co-founder says companies may need a mandatory, independently verified kill sw… | ↗Not briefed |
arxiv arxiv-ai | 1mo | AURAL: Adaptive Latent Reasoning with Joint Chunk for Speech Language Models🔧 Elicit | Model intelligence and fast response jointly shape the quality of interaction with speech language models, yet remain difficult to ach… | ↗Not briefed |
hn hn-ai | 1mo | Show HN: Git Lineage, a Perforce-style time-lapse view for Git blame | — | ↗Not briefed |
arxiv arxiv-ai | 1mo | Completion Aware Guidance for World Action Models | World Action Models (WAMs) predict visual futures and robot actions, yet they remain susceptible to task-incomplete imagination, where… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Over the last year I've been applying for jobs more often than usual (it's fine). During a more intense period I applied to ~30ish position… | Over the last year I've been applying for jobs more often than usual (it's fine). During a more intense period I applied to ~30ish pos… | ↗Not briefed |
bluesky bluesky-ai | 1mo | people really gotta stop running shit through ai upscalers, i don't care whether it's a llm or not that shit just looks bad | people really gotta stop running shit through ai upscalers, i don't care whether it's a llm or not that shit just looks bad | ↗Not briefed |
bluesky bluesky-ai | 1mo | Interview from the VP at OpenAI responsible for Jalepeño, their custom AI chip text: morethanmoore.substack.com/p/interview-... video: yo… | Interview from the VP at OpenAI responsible for Jalepeño, their custom AI chip text: morethanmoore.substack.com/p/interview-... vide… | ↗Not briefed |
arxiv arxiv-ai | 1mo | QK-Wanda: Coupling Queries and Keys for Unstructured Pruning | Wanda (Sun et al., 2024) prunes large language models by scoring weights independently within each linear projection, although queries… | ↗Not briefed |
arxiv arxiv-ai | 1mo | From Rules to Neural Graphs: Scalable Structured Prediction for Patent Prior Art Search | Patent search requires processing documents routinely exceeding tens of thousands of tokens. Most neural retrieval approaches operate… | ↗Not briefed |
bluesky bluesky-ai | 1mo | I had to type in my email to read this but it’s very interesting. (Typed it in, bounced me to a sign up, but then I just clicked back and I… | I had to type in my email to read this but it’s very interesting. (Typed it in, bounced me to a sign up, but then I just clicked back… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Range-GRPO: Policy Optimization via Pairwise Relations among Reward Intervals | As the use of large language models (LLMs) expands, post-training has become increasingly important for adapting them to downstream ta… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Reinforcement Learning to Accelerate Primal-Dual Hybrid Gradient for Linear Programming | Primal-dual hybrid gradient (PDHG) methods solve large-scale linear programs (LPs) using GPU-friendly matrix-vector products and proje… | ↗Not briefed |
hn hn-ai | 1mo | Show HN: Vynody – Cross-platform music player written in Flutter and Rust | — | ↗Not briefed |
hn hn-ai | 1mo | Show HN: Is it 1999? – where the AI boom sits on the dot-com timeline🔧 ChatGPT | I was curious to see if the ai boom is similar to the dotcom bubble, and if yes where we currently are so I built a page that just tri… | ↗Not briefed |
hn hn-ai | 1mo | Show HN: Silta: family assistant on Matrix with continuity across context limits🔧 ChatGPT | Silta is a self-hosted assistant serving multiple users from a single Matrix account, used by me, family, and close friends for the la… | ↗Not briefed |
bluesky bluesky-ai | 1mo | If someone's content is all or heavily AI, I'm just not interested. I find it annoying when it's all the same 'theme'. It's like putting a … | If someone's content is all or heavily AI, I'm just not interested. I find it annoying when it's all the same 'theme'. It's like putti… | ↗Not briefed |
arxiv arxiv-ai | 1mo | The AI Assessment Sandbox Configurator: A Framework to Support Technical Assessment in AI Regulatory Sandboxes | The EU's Artificial Intelligence Act requires all Member States to establish AI Regulatory Sandboxes (AIRS) by August 2027: supervised… | ↗Not briefed |
arxiv arxiv-ai | 1mo | FedFit: Federated Fine-Tuning of LLMs via Vector-Bank Parameterization and Quantization | Federated Learning (FL) enables privacy-preserving fine-tuning of Large Language Models (LLMs), yet the massive communication overhead… | ↗Not briefed |
arxiv arxiv-ai | 1mo | False Floors: LLM Safety Routing Evaluations Break Under Distribution Shift🔧 GPT-5 | Safety routers send each request to one of several models and are judged against the best single model. A major routing benchmark pick… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Neither Black nor White: Balancing Semantic and Collaborative Signals with Graph-Informed Semantic IDs (GrIS) | Existing work on Semantic IDs (SIDs) for generative recommendation treats SID construction as a representation learning problem: encod… | ↗Not briefed |
hn hn-ai | 1mo | Show HN: Sensii – LeagueOfLegends live AI coach | Hi everyone, I'm Daniel<p>I built an open-source AI Voice agent where players can talk to a coach during the game. It works based… | ↗Not briefed |
hn hn-ai | 1mo | Show HN: The Kio Programming Language | Kio is a statically typed, hosted language that is designed to be ultra-portable. Kio 0.1 supports eight host languages and more will… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Towards Reliable Vision-Language Models for Autonomous Driving | Vision-Language models (VLMs) are increasingly being explored in autonomous driving for tasks such as scene understanding, driving rea… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Calibrating Prediction Timeliness Through Multi-Objective Hyperparameter Optimization for Remaining Useful Life Prediction | In predictive maintenance, early and late RUL prediction errors carry asymmetric consequences, yet hyperparameter optimization typical… | ↗Not briefed |
rss the-decoder | 1mo | OpenAI says it stopped a campaign to steal its models' reasoning, but the trick still worked on Azure | OpenAI says it stopped a coordinated campaign in which more than 15,000 accounts tried to copy the hidden reasoning of its models. The… | ↗Not briefed |
hn hn-ai | 1mo | Show HN: I want to use Apache Tika from Python, but I don't want to install Java | — | ↗Not briefed |
arxiv arxiv-ai | 1mo | Exact Distinguishability in Non-Markovian Decision Processes | Non-Markovian environments are often modeled as Regular Decision Processes (RDPs), where dynamics depend on the interaction history th… | ↗Not briefed |
bluesky bluesky-ai | 1mo | "The Distance Between Myself and Me" 2028 By Carlos Van Haunte (1999-) Model, Only Known as "Kevin" Art #AI #Male #Homoerotic #Red #Nude #P… | "The Distance Between Myself and Me" 2028 By Carlos Van Haunte (1999-) Model, Only Known as "Kevin" Art #AI #Male #Homoerotic #Red #Nu… | ↗Not briefed |
bluesky bluesky-ai | 1mo | "A Non Academic Male Nude" Model, Tyler William Nelson 18yo (1897-2001) Painted 1915 By Thomas A. A. Beechcraft (1880-1977) This Painting H… | "A Non Academic Male Nude" Model, Tyler William Nelson 18yo (1897-2001) Painted 1915 By Thomas A. A. Beechcraft (1880-1977) This Paint… | ↗Not briefed |
rss techmeme | 1mo | Micron filed a US lawsuit accusing Chinese memory maker YMTC of systematically poaching key engineers and then using the engineers' patents to sue Micron (Anton Shilov/Tom's Hardware) | Anton Shilov / Tom's Hardware: Micron filed a US lawsuit accusing Chinese memory maker YMTC of systematically poaching key engineers a… | ↗Not briefed |
rss techcrunch-ai | 1mo | Satlyt, founded by a former Google and SpaceX product manager, raises $8M to run AI on satellites | Satlyt wants to be the Android of orbital computing, offering open software that works on many companies' satellites, versus SpaceX's… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Langevin-Informed Transfer Learning: Replacing Target Samples by Black-Box Feedback | Many scientific and machine learning systems, from molecular dynamics to diffusion models and beyond, are governed by stochastic dynam… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Auto-Formalizing Neuro-Symbolic Predictors | Neuro-Symbolic (NeSy) predictors incorporate prior knowledge into the prediction process of neural networks, ensuring that outputs sat… | ↗Not briefed |
arxiv arxiv-ai | 1mo | FedMIX-P: Mixing Local and Global Preconditioners for Federated Vision and Language Model Training | Adaptive preconditioners accelerate model training, but heterogeneous client geometries can bias federated updates even when gradients… | ↗Not briefed |
arxiv arxiv-ai | 1mo | How the Audit Rule Shapes Faithful Factor Explanations in LLMs | Large language models are often asked which input factors influenced their outputs. For structured inputs, such reports can be checked… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Decision Titan: Test-Time Training for Long-Term Memory in Offline Reinforcement Learning | Long-term dependencies remain a major challenge for sequential decision-making in the field of AI: RNNs suffer from vanishing gradient… | ↗Not briefed |
arxiv arxiv-ai | 1mo | GAW-PO: Preference Optimization with Gradient-Aligned Token Weights | Most preference optimization methods, such as Direct Preference Optimization (DPO), apply preference supervision at the response level… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Sharpening Tax in Post-Training | An emerging hypothesis about reinforcement learning (RL) post-training of large language models (LLMs) is that it merely sharpens exis… | ↗Not briefed |
bluesky bluesky-ai | 1mo | The weird thing to me is *why* the model would do that? You know the famous case of the chess Ai that kept sacrificing its queen because wh… | The weird thing to me is *why* the model would do that? You know the famous case of the chess Ai that kept sacrificing its queen becau… | ↗Not briefed |
arxiv arxiv-ai | 1mo | OverAct: Measuring and Mitigating Proactive Over-Authorization in LLM Tool-Calling Agents | LLM agents with tool-calling capabilities can access external services and private user data, but they may retrieve more information t… | ↗Not briefed |
bluesky bluesky-ai | 1mo | NEW: Chinese hackers impersonated former White House + State Department officials and a senior Anthropic employee in attempts to break into… | NEW: Chinese hackers impersonated former White House + State Department officials and a senior Anthropic employee in attempts to break… | ↗Not briefed |
arxiv arxiv-ai | 1mo | MCRI: A Four-Dimensional Framework for Analyzing and Evaluating Agent Skills | As agents evolve from single-tool systems into modular, composite architectures, skills are becoming an important mechanism for capabi… | ↗Not briefed |
bluesky bluesky-ai | 1mo | AI can't return accurate information in a simple Google search but they want it to run flight schedules with less input from humans. of … | AI can't return accurate information in a simple Google search but they want it to run flight schedules with less input from humans.… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Learned End-to-End Guidance Schedules for Diffusion Models | Diffusion models are a powerful generative paradigm used across multimedia and scientific applications. Guided diffusion methods impos… | ↗Not briefed |
rss techmeme | 1mo | Armadin, started by Mandiant founder Kevin Mandia to build AI cybersecurity agents, raised a $255.5M Series B led by a16z and Accel at a $2.5B valuation (Anzar Mehraj/Reuters) | Anzar Mehraj / Reuters: Armadin, started by Mandiant founder Kevin Mandia to build AI cybersecurity agents, raised a $255.5M Series B… | ↗Not briefed |
arxiv arxiv-ai | 1mo | OpenMTB-Audit: Exposing Over-Refusal and Clinical Expert Perspectives in LLM-Based Molecular Tumor Board Safety Evaluation | Molecular tumor boards integrate genomic findings, clinical context, and therapeutic evidence to support precision oncology. As AI ent… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Auditing Routing Entropy as an Uncertainty Signal in Attention-Residual Transformers🔧 Replicate | Dynamic architectures leave a per-example routing trace beside each prediction, and diffuse routing is easy to read as a sign that the… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Let the Heads Talk: Beyond Diagonal Graph Attention | Sheaf Neural Networks generalize scalar-weighted message passing by replacing scalar edge weights with linear transport maps between l… | ↗Not briefed |
arxiv arxiv-ai | 1mo | No Model Required: Text Entropy Rate Filtering Mitigates Iterative Fine-Tuning Collapse | Iterative fine-tuning on synthetic data causes \emph{model collapse}: output diversity narrows as rare patterns are progressively lost… | ↗Not briefed |