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 | 25d | Self-Reflective Chain-of-Thought (SRCoT): Evidence-Based Text Generation via a Closed-Loop Retrieve--Reflect--Revise Pipeline on Edge LLMs | Large language models (LLMs) can generate fluent outputs that lack source support or conflict with local operational constraints. We i… | ↗Not briefed |
bluesky bluesky-ai | 25d | Entendu tout à l'heure : "Claude. ai dit que tu peux faire confiance à son code, que ça ne changera pas le résultat attendu" ... ... ... C… | Entendu tout à l'heure : "Claude. ai dit que tu peux faire confiance à son code, que ça ne changera pas le résultat attendu" ... ...… | ↗Not briefed |
arxiv arxiv-ai | 25d | Inferring L1 Backgrounds from L2 English with LLM Token Probabilities | Native language identification (NLI) infers a non-native speaker's first language (L1) from their second language (L2) production. We… | ↗Not briefed |
arxiv arxiv-ai | 25d | MobileInject: Evaluating Reasoning-to-Action Security of MLLM-Based Mobile Agents under Cross-Modal Consistency Injection🔧 Modal | Multimodal large language model (MLLM)-based mobile agents are rapidly evolving from conversational assistants into autonomous systems… | ↗Not briefed |
arxiv arxiv-ai | 25d | SimOPD: Relational Knowledge Transfer for On-Policy Distillation | A key challenge in on-policy distillation (OPD) is extracting effective supervision from teacher knowledge. Supervision from token pro… | ↗Not briefed |
bluesky bluesky-ai | 25d | There's something deeply resonate about how the AI companies are leveraged up to the gills both metaphorically and literally. Like the enti… | There's something deeply resonate about how the AI companies are leveraged up to the gills both metaphorically and literally. Like the… | ↗Not briefed |
arxiv arxiv-ai | 25d | Designing for Imperfect Prompts: A Simulation-Based Evaluation of an LLM-Based Enterprise Data Assistant Using Knowledge Graphs | Enterprise data assistants using large language models (LLMs) promise democratized data access in organizations. Yet in our real-world… | ↗Not briefed |
arxiv arxiv-ai | 25d | Look, Reason, Draw a Region: Recursive Refinement for Spatial Context Prediction | Spatial context prediction requires visual grounding and spatial reasoning to identify vacant regions relative to objects in a scene.… | ↗Not briefed |
bluesky bluesky-ai | 25d | Backdooring Open-Weight AI Models This is a problem: Any open model whose weights have been edited can carry a backdoor, whether it's a ta… | Backdooring Open-Weight AI Models This is a problem: Any open model whose weights have been edited can carry a backdoor, whether it's… | ↗Not briefed |
arxiv arxiv-ai | 25d | Optimist Deep Research: An Iterative, State-Centric Agent for Autonomous Literature Research🔧 Semantic Scholar | Most retrieval-augmented systems treat literature search as a single retrieval step followed by summarization. Real literature researc… | ↗Not briefed |
arxiv arxiv-ai | 25d | How the elastic axis folds the transonic flutter boundary: shock energy transfer and conformally calibrated active learning | Isogai's case A, a pitch-plunge aerofoil whose elastic axis lies half a chord ahead of the leading edge, has a folded transonic flutte… | ↗Not briefed |
arxiv arxiv-ai | 25d | FlashReFocus: Interactive Image Refocusing in Seconds | Single-image refocusing changes the focal plane and depth of field of a photograph after capture. It is inherently interactive: users… | ↗Not briefed |
arxiv arxiv-ai | 25d | Learning angles for quantum approximate optimization | The Quantum Approximate Optimization Algorithm (QAOA) and its variants are leading algorithms to help solve challenging combinatorial… | ↗Not briefed |
arxiv arxiv-ai | 25d | The Optimal Ranking Tradeoffs Between Two Ranking Scores in Two-Class Classification | The accuracy, true negative rate (negative class recall), true positive rate (positive class recall), negative predictive value (negat… | ↗Not briefed |
bluesky bluesky-ai | 25d | President Donald Trump’s AI task force in its first meeting agreed on Thursday to address risks related to cybersecurity, biosecurity and “… | President Donald Trump’s AI task force in its first meeting agreed on Thursday to address risks related to cybersecurity, biosecurity… | ↗Not briefed |
rss wired-ai | 25d | AI-Generated Hot Girls Are Ripping Off Comedians to Boost Their Horny Content | A network of AI-generated models is using comics’ material to funnel views to their OnlyFans and Fanvue accounts—and there’s not much… | ↗Not briefed |
arxiv arxiv-ai | 25d | Distributed Performance Boosting through Neural Control for Plug-and-Play Networks | We address the problem of improving the performance of pre-stabilized networked systems subject to plug-and-play (PnP) operations, nam… | ↗Not briefed |
arxiv arxiv-ai | 25d | CoTA++: Understanding the Repeat Curse in dMLLMs from an Information Flow Perspective | Diffusion-based multimodal large language models (dMLLMs) decode by iterative denoising with bidirectional attention, which precludes… | ↗Not briefed |
arxiv arxiv-ai | 25d | Data-Efficient Classifier Pretraining Through Diffusion Models | Machine learning with limited data is susceptible to overfitting, especially with modern deep neural networks. Prior to the introducti… | ↗Not briefed |
hn hn-ai | 25d | Unity Mass-Suspends Accounts | Years ago after the last Unity CEO issue, a lot of Unity developers began searching for a new development platform due to pricing chan… | ↗Not briefed |
arxiv arxiv-ai | 25d | Annotrace: A Workbench for Step-Level Annotation of Agent Traces | Every study of why LLM agents fail begins by labelling traces by hand, and every such study builds its own throwaway tool. Observabili… | ↗Not briefed |
rss techmeme | 25d | US candidates and PACs, mostly tied to Republicans, have spent $137M+ on 220 AI-generated campaign ads in 2026; ~70% of the 220 videos didn't disclose AI use (Stuart A. Thompson/New York Times) | Stuart A. Thompson / New York Times: US candidates and PACs, mostly tied to Republicans, have spent $137M+ on 220 AI-generated campaig… | ↗Not briefed |
bluesky bluesky-ai | 25d | 3 dolar kazandım ve hemen sanal kart verdiler . Bakiye yüklendi ve deepseek yatırdım parayı:))))) Para geldi mi geldi Harcadım mı harcadım | 3 dolar kazandım ve hemen sanal kart verdiler . Bakiye yüklendi ve deepseek yatırdım parayı:))))) Para geldi mi geldi Harcadım mı harc… | ↗Not briefed |
bluesky bluesky-ai | 25d | Wrapped zkAPI + Tor + extra cover traffic + local LLM prompt rewriting in an open-AI compatible daemon, and made a "paranoid mode" wrapper … | Wrapped zkAPI + Tor + extra cover traffic + local LLM prompt rewriting in an open-AI compatible daemon, and made a "paranoid mode" wra… | ↗Not briefed |
arxiv arxiv-ai | 25d | Verbalising Model's Computation for Chain-of-Thought Faithfulness Detection | Current Chain-of-Thought (CoT) verification methods detect reasoning faithfulness from model outputs or static activations, but are li… | ↗Not briefed |
bluesky bluesky-ai | 25d | The biggest tell for me is that his response was “ask Grok” You need to be Twitter-brained to heretofore unseen levels to have Grok as the…🔧 ChatGPT | The biggest tell for me is that his response was “ask Grok” You need to be Twitter-brained to heretofore unseen levels to have Grok a… | ↗Not briefed |
arxiv arxiv-ai | 25d | DeDiCa: A Scalable Deconfounding Discrete Causal Generative Model | We introduce DeDiCa, a practical framework for efficient causal inference with discrete variables and hidden confounding that revisits… | ↗Not briefed |
arxiv arxiv-ai | 25d | NestedVLA: Learning to Consolidate and Generate Skills for Vision-Language-Action Models | Vision--Language--Action (VLA) models promise generalist robotic agents, yet consolidating skills learned across heterogeneous tasks a… | ↗Not briefed |
bluesky bluesky-ai | 25d | India's cyber agency says AI agents broke into Indian organizations' systems and read confidential files undetected; AI model providers fla… | India's cyber agency says AI agents broke into Indian organizations' systems and read confidential files undetected; AI model provider… | ↗Not briefed |
rss techmeme | 25d | India's cyber agency says AI agents broke into Indian organizations' systems and read confidential files undetected; AI model providers flagged the incidents (Munsif Vengattil/Reuters) | Munsif Vengattil / Reuters: India's cyber agency says AI agents broke into Indian organizations' systems and read confidential files u… | ↗Not briefed |
hn hn-ai | 25d | Reap the processes your coding agents leave behind | — | ↗Not briefed |
arxiv arxiv-ai | 25d | Beyond Marginal Influence: Interaction-Aware Memory Selection for LLM Agents | LLM-agent memory selection often uses retrieval rankings based on record relevance or individual influence. Such rankings cannot ident… | ↗Not briefed |
arxiv arxiv-ai | 25d | ADT: All-Atom Diffusion Tokenizer for 3D Molecular Structures | Discrete structural tokenization provides a promising route to autoregressive 3D molecular generation. Yet, existing tokenizers face a… | ↗Not briefed |
rss techmeme | 25d | China is considering adding 23 AI technologies to its export control list, including LLMs and chip designs, to keep closer control over homegrown advancements (Nectar Gan/Bloomberg) | Nectar Gan / Bloomberg: China is considering adding 23 AI technologies to its export control list, including LLMs and chip designs, to… | ↗Not briefed |
hn hn-ai | 25d | If a Switch is an ARM64-verse. Then ScarletNX is FEX but in reverse | — | ↗Not briefed |
arxiv arxiv-ai | 25d | Reasoning Models Can Control Their Chains of Thought When They Tell Themselves To | Chain-of-thought (CoT) controllability, the ability of reasoning models to control properties of their CoT, is consequential for both… | ↗Not briefed |
arxiv arxiv-ai | 25d | Beyond Reward Scoring: Online Multi-Agent Preference Learning for Multimodal Response Generation | Multimodal preference alignment often follows a score-based RLHF pipeline: a reward model is first trained to learn human preferences,… | ↗Not briefed |
arxiv arxiv-ai | 25d | Sample, Mix, or Verify? Allocating Test-Time Compute under Imperfect Verifiers | A common way to make language models and agents more accurate is to spend more compute at inference time: generate many candidate solu… | ↗Not briefed |
arxiv arxiv-ai | 25d | Ignored, Not Refused: What Low Prompt-Injection Success Rates Hide in Small Tool-Using Agents | Tool-using agents read untrusted external content, which exposes them to indirect prompt injection (IPI), where hidden instructions hi… | ↗Not briefed |
bluesky bluesky-ai | 25d | 今更ながら登場する3人娘のプロフィールを固定にしました 明日以降はこのプロフィールを使って生まれるためフォルムが変わっちゃう子がいると思いますが大目に見てください笑 ハイブリッドアニメだとあまりにリアル寄りになるのでデジタルアニメ風で← 【 参加者募集中 】 ブログ・日記・… | 今更ながら登場する3人娘のプロフィールを固定にしました 明日以降はこのプロフィールを使って生まれるためフォルムが変わっちゃう子がいると思いますが大目に見てください笑 ハイブリッドアニメだとあまりにリアル寄りになるのでデジタルアニメ風で← 【 参加者募集中 】 ブロ… | ↗Not briefed |
arxiv arxiv-ai | 25d | Preventing Data Hallucinations in Agentic AI for Optical Network Troubleshooting | An agentic AI system for automated troubleshooting in optical communication networks is demonstrated that integrates data references i… | ↗Not briefed |
arxiv arxiv-ai | 25d | FLARE: Physics-Grounded Evaluation of Language-Model Agents for Feeder Restoration with Degraded Observations | Restoring service after distribution-grid faults requires decisions from incomplete and conflicting information while respecting elect… | ↗Not briefed |
arxiv arxiv-ai | 25d | Who Pays for Efficient Inference? Confidence-Based Small-to-Large LLM Routing Across Tasks and Languages | Small-to-large language model routing can reduce inference cost by escalating difficult inputs while retaining easier ones on a smalle… | ↗Not briefed |
arxiv arxiv-ai | 25d | Can the Parts Fool the Test? Counterfactual Pairing Cycles for Relational OOD | Relational out-of-distribution errors are easy to describe and hard to evaluate cleanly. An image may be natural, a caption may be flu… | ↗Not briefed |
bluesky bluesky-ai | 25d | This, from the folks that told employees to use AI or get fired. Sweet baby cheeses. | This, from the folks that told employees to use AI or get fired. Sweet baby cheeses. | ↗Not briefed |
hn hn-ai | 25d | How the Claude Code team does code review | — | ↗Not briefed |
arxiv arxiv-ai | 25d | Think More or Read More? A Token-Budget Theory of Reasoning with Shared Context | Reasoning systems increasingly improve answers by incorporating additional context into the model's input, from multi-agent debate, wh… | ↗Not briefed |
bluesky bluesky-ai | 25d | This is really just a reminder that there is limited ROI in this whole LLM project. The investments will never be paid off. | This is really just a reminder that there is limited ROI in this whole LLM project. The investments will never be paid off. | ↗Not briefed |
hn hn-ai | 25d | Anatomy of a Machine's Ancestors via Claude | — | ↗Not briefed |
arxiv arxiv-ai | 25d | MGE-PO: Entropy-Aware Credit Assignment for Mathematical Reasoning | GRPO assigns the same group-relative advantage to equally rewarded responses within a question, leaving their generation uncertainty u… | ↗Not briefed |