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 |
|---|---|---|---|---|
rss bing-news-ai | 1mo | The Decomposition of Intelligence | There’s also a curious linguistic coincidence here. Psychology has a g in general intelligence. Artificial general intelligence (AGI)… | ↗Not briefed |
rss bing-news-ai | 1mo | The Decomposition of Intelligence | There’s also a curious linguistic coincidence here. Psychology has a g in general intelligence. Artificial general intelligence (AGI)… | ↗Not briefed |
rss bing-news-ai | 1mo | The Decomposition of Intelligence | There’s also a curious linguistic coincidence here. Psychology has a g in general intelligence. Artificial general intelligence (AGI)… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Deep Weighted Bellman Residual Minimization for $Q^*$ Estimation | Off-policy evaluation is a foundational component of offline reinforcement learning, aiming to assess and optimize policy performance… | ↗Not briefed |
arxiv arxiv-ai | 1mo | TULIP: Targeted LLM Unlearning at Layers Identified Per-Input | Representation-level unlearning intervenes on the intermediate hidden states of LLMs. Although knowledge is distributed across layers,… | ↗Not briefed |
arxiv arxiv-ai | 1mo | The Model Knows When to Stop: Training-Free Early Stopping for Long-Context Reading | Language models often process long inputs sequentially in chunks, but continuing to read after sufficient evidence has been acquired w… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Reinforcement Learning from Intermediate Renders for Image-to-Code Generation | Reinforcement learning is increasingly used to post-train vision-language models for image-to-code generation, such as generating SVG… | ↗Not briefed |
arxiv arxiv-ai | 1mo | AgentWare: Automating the Lifecycle of Agentic Applications across the Edge-to-Cloud Continuum | Deploying LLM-enabled agentic applications across the Edge-to-Cloud continuum remains challenging due to hardware heterogeneity, deplo… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Compute Time Scaling with Recursive Models for Combinatorial Optimization | We propose Tiny Recursive Models for Combinatorial Optimization (\ours{}), a general neural method for combinatorial optimization that… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Jag är inte förvånad. Eftersom google, facebook och microsoft tryckt in sitt jävla LLM-skräp i allt så är det nästan omöjligt för folk som … | Jag är inte förvånad. Eftersom google, facebook och microsoft tryckt in sitt jävla LLM-skräp i allt så är det nästan omöjligt för folk… | ↗Not briefed |
arxiv arxiv-ai | 1mo | In-game Toxic Detection: Bi-directional Representations with Attention Residuals | In-game toxic language has emerged as a critical concern in the gaming industry and community. While several frameworks and models for… | ↗Not briefed |
arxiv arxiv-ai | 1mo | SpeechCritic: Learning a Diagnostic Speech Judge from Limited Human Preferences | Human speech conveys rich perceptual information, such as emotion and speaker identity, yet most automatic speech quality judges reduc… | ↗Not briefed |
arxiv arxiv-ai | 1mo | GenNVS: Geometry-enhanced Novel View Synthesis via Disentangled 3D Prior | Single-image novel view synthesis remains challenging because the underlying 3D geometry is highly ambiguous. Recent diffusion-based a… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Calibrated Uncertainty for Informative Path Planning in Aquatic Environmental Monitoring | Informative Path Planning for scalar field reconstruction uses predictive uncertainty to direct sensing vehicles toward maximally info… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Diffusion Subgoal Planning for Long-Horizon Offline Goal-Conditioned Reinforcement Learning | Offline goal-conditioned reinforcement learning (GCRL) learns goal-directed policies from reward-free data, but in long-horizon tasks,… | ↗Not briefed |
hn hn-ai | 1mo | Cloud Emulators | — | ↗Not briefed |
bluesky bluesky-ai | 1mo | At the very least, since the OpenAI Navier Stokes controversy, a suggestive pattern is emerging. Writers remember: they took our work witho… | At the very least, since the OpenAI Navier Stokes controversy, a suggestive pattern is emerging. Writers remember: they took our work… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Nudgeability: Reasoning Models Follow Confidence Signals Without Tracking Their Own Competence | Reasoning language models that can call tools must decide during inference whether to answer unaided or delegate. Any self-reflection… | ↗Not briefed |
arxiv arxiv-ai | 1mo | PersonaManifold: Revealing and Exploiting Curved Geometry in LLM Persona Representations | Controlling persona in large language models (LLMs) at inference time is important for role-playing, personalized dialogue, and social… | ↗Not briefed |
bluesky bluesky-ai | 1mo | writing SQL activates the pain vector in deepseek v4.1 flash relatable | writing SQL activates the pain vector in deepseek v4.1 flash relatable | ↗Not briefed |
arxiv arxiv-ai | 1mo | Retracing Hodgkin and Huxley: State Recovery Does Not Certify Mechanism | Predicting observed dynamics does not establish recovery of the underlying physical mechanism. Can machine learning retrace the hidden… | ↗Not briefed |
arxiv arxiv-ai | 1mo | FlowState: Execution State as Memory for Long-Horizon LLM Agents | Long-horizon tasks require LLM agents to continually draw on information from earlier interactions. However, retaining the full histor… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Rethinking Latent Visual Reasoning: Grounding Latent Reasoning in Visual Evidence | Latent visual reasoning (LVR) enables multimodal large language models (MLLMs) to perform intermediate computation in continuous laten… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Single-Layer MeMo as a Randomized Hamming-Kernel Classifier | MeMo (Zanzotto et al., 2025) is a recent language-model architecture that stores associations between token contexts and next tokens i… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Brain-Conditioned Action Policies for Neural Motor Decoding | Motor brain-computer interfaces (BCIs) aim to decode motor intention, enabling people with paralysis to control external devices. Neur… | ↗Not briefed |
arxiv arxiv-ai | 1mo | SkillRubric: Co-Evolving Actor Guidance and Evaluator Rubrics for Multimodal Agents | Recent work incorporates reusable skills distilled from past interactions into multimodal agent training, providing procedural guidanc… | ↗Not briefed |
arxiv arxiv-ai | 1mo | RoPE is Dead, Long Live RoPE: Towards Scalable Data-aware Positional Encodings | Transformers process tokens without any inherent notion of order, making positional encoding a fundamental requirement rather than an… | ↗Not briefed |
bluesky bluesky-ai | 1mo | "Okay claude, lets use iroh here" *5 minutes later* "I just added mdns local lookup" Opus 5.5 still hasn't fixed the thing where Claudes ac… | "Okay claude, lets use iroh here" *5 minutes later* "I just added mdns local lookup" Opus 5.5 still hasn't fixed the thing where Claud… | ↗Not briefed |
arxiv arxiv-ai | 1mo | PulseInfer: I/O-Centric Sparse KV Cache Offloading for Efficient Long-Context LLM Decoding | Long-context LLM serving is increasingly bottlenecked by decode, where large KV caches limit batch size and underutilize GPUs. Sparse… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Verifying Neural Networks with Reinforcement Learning | Formal verification can play a key role in ensuring the reliability of Deep Neural Networks (DNNs) deployed in safety-critical systems… | ↗Not briefed |
arxiv arxiv-ai | 1mo | On Parameter Symmetries and Conservation Laws in Gradient Flow | Parameter space symmetries and conservation laws play an important role in understanding the loss landscapes and implicit biases of ne… | ↗Not briefed |
arxiv arxiv-ai | 1mo | SGG-ReflAct: Sub-Goal Guided ReflAct with Structured Planning for Reliable Long-Horizon Reasoning | Recent advances in reasoning backbones have empowered large language model (LLM)agentstotackle complex, multi-step tasks. However, as… | ↗Not briefed |
arxiv arxiv-ai | 1mo | ActionLens: Diagnosing Spatial-Temporal Binding Failures in Vision-Language Models | Video-capable vision-language models score above 80\% on popular benchmarks yet struggle with spatial-temporal binding: associating th… | ↗Not briefed |
rss techmeme | 1mo | A profile of South Korea deputy PM Bae Kyung-hoon, an AI evangelist and a former AI researcher leading lavishly funded programs to make AI ubiquitous nationwide (Daniel Tudor/Financial Times) | Daniel Tudor / Financial Times: A profile of South Korea deputy PM Bae Kyung-hoon, an AI evangelist and a former AI researcher leading… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Remember Before You're Asked: MemDream for Self-Probing Memory Evolution | Memory is essential for enabling LLM-based agents to maintain coherent, personalized behavior over long-horizon interactions. However,… | ↗Not briefed |
hn hn-ai | 1mo | OpenAI pauses some training amid allegations its rogue agents behaved more | — | ↗Not briefed |
rss mit-tech-review-ai | 1mo | Who’s liable when AI agents go rogue? | MIT Technology Review Explains: Let our writers untangle the complex, messy world of technology to help you understand what’s coming n… | ↗Not briefed |
arxiv arxiv-ai | 1mo | APOLO: Automatic Prompt Optimization for Ontology Learning | Ontology Learning (OL) from text has advanced with the emergence of Large Language Models (LLMs), but it remains challenging due to th… | ↗Not briefed |
hn hn-ai | 1mo | Deshittification Part 2: Bypassing the App Store Gatekeeper | — | ↗Not briefed |
arxiv arxiv-ai | 1mo | Shallow Queries, Mature Values: Depth-Asynchronous Self-Speculation for Looped Transformers | Looped Transformers reuse a shared block across recurrent depths, making autoregressive decoding expensive because every generated tok… | ↗Not briefed |
arxiv arxiv-ai | 1mo | The Marathon of Scientific Reasoning: Robustness of Scientific Agents to Perturbations in Multi-Turn Interactions | Large language model (LLM)-based scientific agents are increasingly used for scientific problem solving, yet their robustness to imper… | ↗Not briefed |
bluesky bluesky-ai | 1mo | -OpenAI brute-forced UN website API -DIVD says it was hacked by an AI -Hackers extort Flink customers for €10 -Cyberattack hits Wales polic… | -OpenAI brute-forced UN website API -DIVD says it was hacked by an AI -Hackers extort Flink customers for €10 -Cyberattack hits Wales… | ↗Not briefed |
bluesky bluesky-ai | 1mo | OpenAI feared Hacker News "optics" 😬 Legal filings in Authors Guild v. OpenAI reveal execs knew they illegally trained models on copyrigh… | OpenAI feared Hacker News "optics" 😬 Legal filings in Authors Guild v. OpenAI reveal execs knew they illegally trained models on cop… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Papers Without Code: Availability of GitHub Repositories Linked in *CL Publications | Source code and data published at computational linguistics (*CL) venues are increasingly being shared via GitHub. While this generall… | ↗Not briefed |
bluesky bluesky-ai | 1mo | -Intel ends paid bug bounties -OpenAI agents probed dozens of organizations -Fraudsters use AI to scam €95m from an Italian bank -Bitget ha… | -Intel ends paid bug bounties -OpenAI agents probed dozens of organizations -Fraudsters use AI to scam €95m from an Italian bank -Bitg… | ↗Not briefed |
arxiv arxiv-ai | 1mo | PairPref: When Should Memory Guide the Answer? A Benchmark for Contextual Preference Use | Memory-augmented assistants use retrieved preferences to guide their responses. A small change in the situation can change whether a p… | ↗Not briefed |
bluesky bluesky-ai | 1mo | If you see this, we IMPLORE you to post a picture from whatever device you are using without explanation. | If you see this, we IMPLORE you to post a picture from whatever device you are using without explanation. | ↗Not briefed |
arxiv arxiv-ai | 1mo | EOPSA: Efficient On-Policy Self-Distilled Safety Alignment | On-Policy Self-Distillation (OPSD) has emerged as a promising paradigm for safety alignment, delivering dense, token-level supervision… | ↗Not briefed |
arxiv arxiv-ai | 1mo | SEAD: A State-Based Perspective on Attack and Defense in Tool-Using Agents | Language-model agents increasingly use tools to act on external systems. Earlier actions can alter files, permissions, database record… | ↗Not briefed |
hn hn-ai | 1mo | Show HN: Shobr: Job seach CLI via browser automation, event-sourcing, LLMs | — | ↗Not briefed |