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 | 29d | I liked this article better when I wrote it. newsdeskdaily.com/latest-news/... I wonder if they could put quotation marks around the half… | I liked this article better when I wrote it. newsdeskdaily.com/latest-news/... I wonder if they could put quotation marks around the… | ↗Not briefed |
arxiv arxiv-ai | 29d | PN-Mix: Pair-Norm Guided Mixed-Precision Key Cache Quantization for Efficient Long-Output LLM Inference | Long-output inference in large language models suffers from severe KV cache memory and bandwidth bottlenecks. While KV cache quantizat… | ↗Not briefed |
hn hn-ai | 29d | Neurochat Running on Community GPUs | — | ↗Not briefed |
bluesky bluesky-ai | 29d | back in my day, spies had to go to spy school to learn how to build spy tech; now, any tom, dick, and harry can go to openai dot com and ge… | back in my day, spies had to go to spy school to learn how to build spy tech; now, any tom, dick, and harry can go to openai dot com a… | ↗Not briefed |
hn hn-ai | 29d | Show HN: Geometry Check, a Fusion 360 add-in that finds slightly-wrong geometry | — | ↗Not briefed |
arxiv arxiv-ai | 29d | Beyond Static Positional Encoding: Bayesian Time-aware Attention for Non-Stationary Long-term Time Series Forecasting | Transformers have demonstrated strong performance in long-term time series forecasting (LTSF), yet modeling non-stationary temporal de… | ↗Not briefed |
arxiv arxiv-ai | 29d | PA1-OPD: One-Step Prefix-Augmented On-Policy Distillation for Language Model Pretraining | Knowledge distillation (KD) improves language-model pretraining by matching teacher distributions at corpus prefixes, but provides no… | ↗Not briefed |
bluesky bluesky-ai | 29d | Пишут, интеллектуально-квестовая составляющая «Бегущего Города» в эпоху LLM с кастомными харнессами таки достигла идеала, предсказанного го… | Пишут, интеллектуально-квестовая составляющая «Бегущего Города» в эпоху LLM с кастомными харнессами таки достигла идеала, предсказанно… | ↗Not briefed |
bluesky bluesky-ai | 29d | Yup... Many (serious) people are saying... (ICYMI www.di.ens.fr/~fbach/MJord...) | Yup... Many (serious) people are saying... (ICYMI www.di.ens.fr/~fbach/MJord...) | ↗Not briefed |
arxiv arxiv-ai | 29d | JevCoder: When Does Decision Delegation Pay Off in Code Repair? | A coding agent repeatedly makes decisions as it repairs code. These decisions shape repair trajectories and affect the token cost of s… | ↗Not briefed |
bluesky bluesky-ai | 29d | どこもかしこもサイバーテロまみれの日常になってしまったなぁ… マジOpenAI責任とれよほんと | どこもかしこもサイバーテロまみれの日常になってしまったなぁ… マジOpenAI責任とれよほんと | ↗Not briefed |
arxiv arxiv-ai | 29d | InstructPro2: Natural Language and Biological Grounding Guided De Novo Protein Design | The ability to generate proteins with desired functions is a central goal of computational biology. Although natural language can expr… | ↗Not briefed |
hn hn-ai | 29d | PS5LM: Llama.cpp on a jailbroken PS5, run any GGUF model natively on the console | — | ↗Not briefed |
rss marktechpost | 29d | Google Releases Gemini 4 Argon API With 1M Token Output Limit | Google has made Gemini 4 Argon generally available on the Gemini API, AI Studio and the Gemini app. The sparse MoE model writes up to… | ↗Not briefed |
arxiv arxiv-ai | 29d | L3M: A Verifiable Harness and Runtime for Self-Extending and Self-Improving Agents | The lasting gains in AI have come from methods that absorb more computation, and language models already scale computation within a ca… | ↗Not briefed |
bluesky bluesky-ai | 29d | The battle over the AI-led future of banking has begun ft.trib.al/rAjkc50 | opinion | The battle over the AI-led future of banking has begun ft.trib.al/rAjkc50 | opinion | ↗Not briefed |
bluesky bluesky-ai | 29d | In einem Konferenzraum von OpenAI wurde ein verstecktes Aufzeichnungsgerät entdeckt. Der Mann, der es angebracht haben soll, wurde festgeno… | In einem Konferenzraum von OpenAI wurde ein verstecktes Aufzeichnungsgerät entdeckt. Der Mann, der es angebracht haben soll, wurde fes… | ↗Not briefed |
bluesky bluesky-ai | 29d | this girl in jinhua asked if i had used deepseek and she was so hype when i said yes lol | this girl in jinhua asked if i had used deepseek and she was so hype when i said yes lol | ↗Not briefed |
arxiv arxiv-ai | 29d | JockeyBench: Evaluating LLM Agents on Horse-Racing Decision-Making | Repeated event forecasting requires extracting predictive signals from noisy evidence and evaluating their value across successive dec… | ↗Not briefed |
bluesky bluesky-ai | 29d | one of the worst things about big monorepos is that to do them right you imo need approvals and extensive CI pipelines which is why they sh… | one of the worst things about big monorepos is that to do them right you imo need approvals and extensive CI pipelines which is why th… | ↗Not briefed |
arxiv arxiv-ai | 29d | Projected Wasserstein-Fisher-Rao Flows for Hawkes Background-Intensity Estimation | We consider the problem of estimating the background intensity of a Hawkes process when the triggering mechanism is known, treating th… | ↗Not briefed |
arxiv arxiv-ai | 29d | Moral Combat: Exploiting Moral Foundations Theory to Jailbreak LLMs | Social psychology has long described how people can be persuaded to interpret a harmful act as moral. Recent research suggests that La… | ↗Not briefed |
bluesky bluesky-ai | 29d | お昼は温かい味噌汁に九州のささがきゴボウも入れました☺️ 私は読書中😅 おババ様が言うには安全性を第一にAnthropicのAIは構築されAmazon、Googleが資金を提供していると···✌️ アンソロピック社はNTTと日本大阪にデータセンター？を作る？？ アメリカのG… | お昼は温かい味噌汁に九州のささがきゴボウも入れました☺️ 私は読書中😅 おババ様が言うには安全性を第一にAnthropicのAIは構築されAmazon、Googleが資金を提供していると···✌️ アンソロピック社はNTTと日本大阪にデータセンター？を作る？？ ア… | ↗Not briefed |
bluesky bluesky-ai | 29d | Oh wat een schitterende ironie, OpenAI is wél bezorgd om hun eigen privacy en gegevens | Oh wat een schitterende ironie, OpenAI is wél bezorgd om hun eigen privacy en gegevens | ↗Not briefed |
arxiv arxiv-ai | 29d | Who's Behind the Text? Robust Model Authorship Attribution | As large language models (LLMs) are increasingly used to generate text, their misuse has become a growing societal problem, with impli… | ↗Not briefed |
arxiv arxiv-ai | 29d | XferAgent: Cross-Agent Experience Transfer for Self-Improving LLM Agents | Large language model (LLM) agents can reuse experience from prior task executions to improve performance on subsequent tasks. Although… | ↗Not briefed |
bluesky bluesky-ai | 29d | oh, and you can start using our product soon. the thing that anthropic build and won’t release? we built something similar, but you can use… | oh, and you can start using our product soon. the thing that anthropic build and won’t release? we built something similar, but you ca… | ↗Not briefed |
bluesky bluesky-ai | 29d | Anthropic plug in for law students. *stares into the distance* | Anthropic plug in for law students. *stares into the distance* | ↗Not briefed |
rss techmeme | 29d | US defense companies and startups have signed at least 13 deals with Taiwanese hardware companies since September 2025 to build a secure drone tech supply chain (Joyu Wang/Wall Street Journal) | Joyu Wang / Wall Street Journal: US defense companies and startups have signed at least 13 deals with Taiwanese hardware companies sin… | ↗Not briefed |
bluesky bluesky-ai | 29d | ME: "So, can the new Muse Wonder collect Facebook data?" AI: "You can't use Claude for that." ME: "Wait you just said Claude where did that… | ME: "So, can the new Muse Wonder collect Facebook data?" AI: "You can't use Claude for that." ME: "Wait you just said Claude where did… | ↗Not briefed |
arxiv arxiv-ai | 29d | Bridging the Topological Divide: Dual-Branch Graph Distillation with Semantic Priors for Robust Cold-Start Recommendation🔧 Modal | Graph Collaborative Filtering models have achieved remarkable success in recommendation systems. However, they inherently suffer from… | ↗Not briefed |
arxiv arxiv-ai | 29d | Learning Across Temporal Graph Representation Silos | Temporal graph data are recorded in different representations, but whether these differences also determine the scope of learned knowl… | ↗Not briefed |
bluesky bluesky-ai | 29d | when I write OpenAI in lower case it feels like Oppai and how maybe openai is an unsavoury verb in Japanese like openai / openasen | when I write OpenAI in lower case it feels like Oppai and how maybe openai is an unsavoury verb in Japanese like openai / openasen | ↗Not briefed |
bluesky bluesky-ai | 29d | 「Gemini 4 Argon」を正式リリース、Claude Fable 5.1やGPT-6 Astraと同等性能で久しぶりにGoogleがAIトップ競争に追いつく https://gigazine.net/news/20261014-gemini-4-argon-availa… | 「Gemini 4 Argon」を正式リリース、Claude Fable 5.1やGPT-6 Astraと同等性能で久しぶりにGoogleがAIトップ競争に追いつく https://gigazine.net/news/20261014-gemini-4-argon-a… | ↗Not briefed |
arxiv arxiv-ai | 29d | On Inherent Privacy of Posterior Sampling: A Unified Rényi-Divergence Framework | Differential Privacy (DP) provides a rigorous framework for quantifying privacy guarantees. While most methods achieve DP by injecting… | ↗Not briefed |
arxiv arxiv-ai | 29d | Divergence-Routed On-Policy Distillation for Language Models | On-policy distillation (OPD) transfers knowledge from a teacher model to a student along student-generated trajectories, typically by… | ↗Not briefed |
bluesky bluesky-ai | 29d | AI models like Claude are turning classic Nintendo games into PC ports overnight! Industrial reverseengineering is shaking the retro scene.… | AI models like Claude are turning classic Nintendo games into PC ports overnight! Industrial reverseengineering is shaking the retro s… | ↗Not briefed |
hn hn-ai | 29d | Show HN: Fabee, an AI nutrition coach, get advice that fits you.(iOS) | — | ↗Not briefed |
bluesky bluesky-ai | 29d | I want to be explicit that Mark Zuckerberg, on the softest of softball podcasts, admitted he invested all this capex without really knowing… | I want to be explicit that Mark Zuckerberg, on the softest of softball podcasts, admitted he invested all this capex without really kn… | ↗Not briefed |
arxiv arxiv-ai | 29d | Geometric Kernel Diversity: An Axiomatic Diversity Measure in Euclidean Space | Quantifying the diversity of a set of objects is a fundamental problem in machine learning and statistics. Mironov and Prokhorenkova (… | ↗Not briefed |
arxiv arxiv-ai | 29d | INTACT: Inducing Latent Reasoning on a Frozen Backbone | Latent reasoning replaces natural-language chains of thought with compact continuous "soft-token" representations, reducing the number… | ↗Not briefed |
bluesky bluesky-ai | 29d | どうすんのこれ Anthropic、小型AIの自己複製に警鐘 数カ月以内に家庭内機器へ拡散も www.digitaltoday.co.kr/jp/view/1140... ＞Anthropicは12日に公開したブログで、AIモデルがスマートオーブンやAndroid TVボック… | どうすんのこれ Anthropic、小型AIの自己複製に警鐘 数カ月以内に家庭内機器へ拡散も www.digitaltoday.co.kr/jp/view/1140... ＞Anthropicは12日に公開したブログで、AIモデルがスマートオーブンやAndroid… | ↗Not briefed |
bluesky bluesky-ai | 29d | AnthropicがGit代替のバージョン管理システム「Scion」を開発し社内移行済み、AIでコードが爆増してGitの限界に達したので耐えられるシステムを開発 https://gigazine.net/news/20261014-anthropic-scion-version… | AnthropicがGit代替のバージョン管理システム「Scion」を開発し社内移行済み、AIでコードが爆増してGitの限界に達したので耐えられるシステムを開発 https://gigazine.net/news/20261014-anthropic-scion-ve… | ↗Not briefed |
bluesky bluesky-ai | 29d | #IA « Les mathématiciens face à l'IA » www.arte.tv/fr/videos/11... #mathématique #sciences #LLM | #IA « Les mathématiciens face à l'IA » www.arte.tv/fr/videos/11... #mathématique #sciences #LLM | ↗Not briefed |
bluesky bluesky-ai | 29d | telling someone that they shouldn't use a 3d model because it almost makes them as bad as someone using AI to generate stuff is like tellin… | telling someone that they shouldn't use a 3d model because it almost makes them as bad as someone using AI to generate stuff is like t… | ↗Not briefed |
arxiv arxiv-ai | 29d | Frame Evidence:A Benchmark for Reference Frames and Visual Evidence in Multi-Image Spatial Reasoning | A spatial answer depends on the reference frame specified in language and its supporting visual evidence. Item accuracy alone obscures… | ↗Not briefed |
arxiv arxiv-ai | 29d | REWIND: Uncertainty-Guided Trajectory Repair for Test-Time Reasoning | An early mistake in the generated reasoning can constrain all subsequent reasoning. Repeated-generation methods, like MCMC-based appro… | ↗Not briefed |
bluesky bluesky-ai | 29d | #MaxTegmark #SafeAILeader is who they should all be consulting, following his lead for #ProHumanAI (if any at all) @futureoflife.org … | #MaxTegmark #SafeAILeader is who they should all be consulting, following his lead for #ProHumanAI (if any at all) @futureoflife.… | ↗Not briefed |
arxiv arxiv-ai | 29d | $τ$-Emotion-Bench: Exposing the Interaction Fragility of LLM Agents through Real-World User Emotions | In long-horizon tool-calling tasks, LLM agents can fail not only because of tool-use errors, but also because of ineffective interacti… | ↗Not briefed |
hn hn-ai | 29d | Show HN: AgentGate – Fast authorization gateway for AI agent tools | — | ↗Not briefed |