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 | Evidence Before Sampling: Interpretable Implicit Negative Candidate Discovery for Recommendation | Recommender systems learn from observed user-item interactions, but explicit negative feedback is often unavailable. Since deep learni… | ↗Not briefed |
arxiv arxiv-ai | 1mo | AgentMemGate: Addressing Speculation Contamination in Conversational Assistant Memory | Conversational AI assistants with long-term memory extract facts from user messages into a store consulted in later conversations. A s… | ↗Not briefed |
arxiv arxiv-ai | 1mo | WASD: Wasserstein-based Knowledge Distillation for Large Language Models | Autoregressive large language models (LLMs) have rapidly advanced in capability, but their increasing scale comes with substantial com… | ↗Not briefed |
arxiv arxiv-ai | 1mo | What Frame-Level Labels Can and Cannot Do for Small-UAV Point Detection in Thermal Video | The growing use of unmanned aerial vehicles (UAVs) has increased the importance of image-based UAV detection. Learning-based detectors… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Oh, not sure. It wrote it's proof for lean, couldn't have been done using agents? What's the argument for diff llm here? | Oh, not sure. It wrote it's proof for lean, couldn't have been done using agents? What's the argument for diff llm here? | ↗Not briefed |
hn hn-ai | 1mo | Windows Storage Proxy Driver – User mode disk storage | — | ↗Not briefed |
bluesky bluesky-ai | 1mo | 仕事で事業会社のDXに関わってきたけど、今までやってきたプロジェクトのほとんどは大企業のおもちゃだと思っている LLMだけはマジで本物 | 仕事で事業会社のDXに関わってきたけど、今までやってきたプロジェクトのほとんどは大企業のおもちゃだと思っている LLMだけはマジで本物 | ↗Not briefed |
bluesky bluesky-ai | 1mo | AIでコード生成は加速、でも人間の確認が追いつかない問題。t-wadaが示す「レビュー解体」という答え type.jp/et/feature/3... を拝読しました。 t-wadaさんが仰ることは、ごもっともなのですが、最後の「人間の理解」の定義が曖昧かな、と感じました。 … | AIでコード生成は加速、でも人間の確認が追いつかない問題。t-wadaが示す「レビュー解体」という答え type.jp/et/feature/3... を拝読しました。 t-wadaさんが仰ることは、ごもっともなのですが、最後の「人間の理解」の定義が曖昧かな、と感じ… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Independent Multi-Agent Reinforcement Learning with Counterfactual Semantic-Social World Models | Fully decentralized multi-agent reinforcement learning (MARL), also referred to as independent learning, requires each agent to learn… | ↗Not briefed |
bluesky bluesky-ai | 1mo | www.commondreams.org/news/sam-alt... “We have to crush these sociopaths,” said one critic. French-style Revolution should include Billion… | www.commondreams.org/news/sam-alt... “We have to crush these sociopaths,” said one critic. French-style Revolution should include Bi… | ↗Not briefed |
rss techmeme | 1mo | Singapore-based data center operator DayOne files for a US IPO, reporting its H1 revenue more than tripled YoY to $512M while its net loss widened to $77.2M (Pragyan Kalita/Reuters) | Pragyan Kalita / Reuters: Singapore-based data center operator DayOne files for a US IPO, reporting its H1 revenue more than tripled Y… | ↗Not briefed |
bluesky bluesky-ai | 1mo | burying the lede Claude, an AI model developed by Anthropic, discovered the algorithm that refutes the 3SUM, APSP, and Exact Triangle hypo… | burying the lede Claude, an AI model developed by Anthropic, discovered the algorithm that refutes the 3SUM, APSP, and Exact Triangle… | ↗Not briefed |
bluesky bluesky-ai | 1mo | also, every time he does videos on this the comments are filled with llm lovers going on these insane essay length tirades about how he’s w… | also, every time he does videos on this the comments are filled with llm lovers going on these insane essay length tirades about how h… | ↗Not briefed |
bluesky bluesky-ai | 1mo | >金にならないからインフラとか狙われないだけで、オーストラリアでは健保情報侵入みたいな先例がある アメリカなんかどうやったのか水道局の制御システムに入り込んでロックして身代金要求する事件とか起きてる。(しかもこれはLLM禍前) 物理的にネットから遮断しろ(専用物理回線を導入… | >金にならないからインフラとか狙われないだけで、オーストラリアでは健保情報侵入みたいな先例がある アメリカなんかどうやったのか水道局の制御システムに入り込んでロックして身代金要求する事件とか起きてる。(しかもこれはLLM禍前) 物理的にネットから遮断しろ(専用物理… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Speaking at a hearing before Australia’s Parliament on Tuesday, OpenAI’s chief strategy officer, Jason Kwon, apologized for the company’s A… | Speaking at a hearing before Australia’s Parliament on Tuesday, OpenAI’s chief strategy officer, Jason Kwon, apologized for the compan… | ↗Not briefed |
arxiv arxiv-ai | 1mo | On the Boundary of Admission Gates: An Injected-Truth Study of Falsification-First Selection in Quantitative Strategy Research | Strategy research conflates two problems: finding a profitable rule, and establishing that the finding is not search luck. The latter… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Detecting LLM-Assisted Vietnamese Writing via Keystrokes under Behavioral Manipulation | We study the robustness of keystroke dynamics for detecting large language model (LLM)-assisted writing. We introduce a Vietnamese key… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Improving Synthetic Data Generation for Argument Mining via Adversarial Reinforcement Learning | Argument Mining (AM) is fundamentally constrained by the scarcity of high-quality structure-annotated datasets. While LLMs have shown… | ↗Not briefed |
bluesky bluesky-ai | 1mo | idk how to be normal about this, people this powerful speaking like this get me thinking of cleansing holy fire, stakes, heresy against lif… | idk how to be normal about this, people this powerful speaking like this get me thinking of cleansing holy fire, stakes, heresy agains… | ↗Not briefed |
bluesky bluesky-ai | 1mo | "Pentagon stops using Anthropic AI tools after blacklisting company, BBC told" www.bbc.com/news/article... | "Pentagon stops using Anthropic AI tools after blacklisting company, BBC told" www.bbc.com/news/article... | ↗Not briefed |
bluesky bluesky-ai | 1mo | LLMは所詮インターネットサービスなんだから普通に法適用内で処分できるのに、それをやろうとしない企業と政府がただ怠惰なだけなんだよな。 最初からLLMに人格も意志もねえよ。 | LLMは所詮インターネットサービスなんだから普通に法適用内で処分できるのに、それをやろうとしない企業と政府がただ怠惰なだけなんだよな。 最初からLLMに人格も意志もねえよ。 | ↗Not briefed |
bluesky bluesky-ai | 1mo | www.commondreams.org/news/sam-alt... | www.commondreams.org/news/sam-alt... | ↗Not briefed |
bluesky bluesky-ai | 1mo | This really is a good LLM slop project if it’s just for personal use, I’m thinking of scraping some recipe sites and serving them out of a … | This really is a good LLM slop project if it’s just for personal use, I’m thinking of scraping some recipe sites and serving them out… | ↗Not briefed |
hn hn-ai | 1mo | Show HN: Counterbranch – See how your PR changes access control | Hi, we just released Counterbranch as a GitHub Action (alpha) and are looking for feedback.<p>The action scans your codebase looking f… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Microsoft internal Claude spend estimate has been slashed by more than a third Meta is experiencing a similar internal shift...Claude Code… | Microsoft internal Claude spend estimate has been slashed by more than a third Meta is experiencing a similar internal shift...Claude… | ↗Not briefed |
bluesky bluesky-ai | 1mo | I need to finish my work so I can break an LLM and have it red team the chat bot they shoved into my property management software | I need to finish my work so I can break an LLM and have it red team the chat bot they shoved into my property management software | ↗Not briefed |
arxiv arxiv-ai | 1mo | BluffJAX: Adversarial Imperfect Information Games in JAX | We introduce BluffJAX: an open-source suite of adversarial imperfect information games in JAX. We provide canonical implementations of… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Disentangling Dual Image References in Frequency Aware Diffusion Models for Personalized Generation | Personalized image generation aims to synthesize text-driven images conditioned on reference images, while mainly casting the generati… | ↗Not briefed |
arxiv arxiv-ai | 1mo | EigenDEXplore: Structured Exploration for Dexterous Manipulation with Human Priors | Dexterous manipulation poses a challenging high-dimensional optimization problem, as useful behaviors require coordinated motion acros… | ↗Not briefed |
bluesky bluesky-ai | 1mo | つまりLLMを語っている人たちの多くが、これが企業の収益を目的としたただの製品・サービスであることが頭からすっかり抜け落ちてる。LLMを正しく評価なんて土台無理だよ | つまりLLMを語っている人たちの多くが、これが企業の収益を目的としたただの製品・サービスであることが頭からすっかり抜け落ちてる。LLMを正しく評価なんて土台無理だよ | ↗Not briefed |
bluesky bluesky-ai | 1mo | LLMと関連性が疑われる不正アクセスはそもそもテック企業が投資家アピールのためスコアにかまけて自社製品殆ど対策をしていなかったことに起因しているんだから、本来ならテック企業に（説明）責任を求めるのが筋だと思うんだけど、メディアも国も及び腰だし、その影響か、国民自体もテック企業に… | LLMと関連性が疑われる不正アクセスはそもそもテック企業が投資家アピールのためスコアにかまけて自社製品殆ど対策をしていなかったことに起因しているんだから、本来ならテック企業に（説明）責任を求めるのが筋だと思うんだけど、メディアも国も及び腰だし、その影響か、国民自体もテ… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Adaptive Model Inversion Attacks Generalize a Privacy-Robustness Tradeoff | In this paper, we show that standard evaluations of high-resolution Model Inversion Attacks (MIAs) significantly underestimate trainin… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Exact-Solution Volume and Length Generalization in Transformers | Research on transformer expressivity shows whether a transformer is capable of solving a given task, but gives little indication of wh… | ↗Not briefed |
arxiv arxiv-ai | 1mo | EIO-Agents: The Missing Semantic Layer for AI Agent Evaluation | AI agents are entering production in increasingly consequential environments without a shared semantic standard for what their evaluat… | ↗Not briefed |
rss techmeme | 1mo | Sources: Moonshot AI closes its final private funding round at a ~$50B valuation and is targeting a Hong Kong IPO in Q1 2027 to raise up to $5B (Bloomberg) | Bloomberg: Sources: Moonshot AI closes its final private funding round at a ~$50B valuation and is targeting a Hong Kong IPO in Q1 202… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Evaluating human-AI workflows for field research in viticulture | We assessed the value of two live human-AI interactions in a precision disease control project in California vineyards. The project te… | ↗Not briefed |
arxiv arxiv-ai | 1mo | CACHEFORGE: LLM-Guided End-to-End Generative Cache Replacement Policy for Performance and Hardware Efficiency | Modern cache replacement designs saturate because they operate within fixed representational structures, hand-crafted and heuristic ba… | ↗Not briefed |
arxiv arxiv-ai | 1mo | SENSE: State-aware Emotion Navigation Storytelling Engine | This paper presents SENSE, a state-aware framework for generating playable branching visual novels with multi-track emotional navigati… | ↗Not briefed |
hn hn-ai | 1mo | LLMs may have immensely helped my RSI | — | ↗Not briefed |
hn hn-ai | 1mo | Iseberg – The PowerShell ISE Reimagined with Avalonia and PowerShell 7 | — | ↗Not briefed |
bluesky bluesky-ai | 1mo | "'The great thing about using AI to check medical information is that it shows RFK Jr. doesn’t know what he is talking about.'" | "'The great thing about using AI to check medical information is that it shows RFK Jr. doesn’t know what he is talking about.'" | ↗Not briefed |
bluesky bluesky-ai | 1mo | South Korea plans to develop $3.5 billion frontier AI model starting next year reut.rs/4hwouip | South Korea plans to develop $3.5 billion frontier AI model starting next year reut.rs/4hwouip | ↗Not briefed |
arxiv arxiv-ai | 1mo | Joint Workflow and Prompt Optimization for User Behavior Simulation | User behavior simulation is the computational modeling of user interactions within information systems through the use of simulated ag… | ↗Not briefed |
arxiv arxiv-ai | 1mo | MS-ECG-FM: Towards a More Universal Electrocardiogram Foundation Model for Health Monitoring using Multi-source Contrastive Learning | Electrocardiography (ECG) records the electrical activity of the heart, aiding diagnosis by detecting abnormalities in cardiac functio… | ↗Not briefed |
bluesky bluesky-ai | 1mo | American AI companies' performance lead over China has narrowed sharply in past months to a record low of 3% after labs such as DeepSeek ga… | American AI companies' performance lead over China has narrowed sharply in past months to a record low of 3% after labs such as DeepSe… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Massive Activation Gating Channel in Large Language Models | Massive activations, a phenomenon in which a small number of hidden channels exhibit exceptionally large magnitudes, are pervasive in… | ↗Not briefed |
arxiv arxiv-ai | 1mo | DLoop: Looped Speculative Decoding | Speculative decoding accelerates autoregressive generation in large language models. In each drafting stage, a lightweight draft model… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Where Rules End and Judges Begin: Measuring the Judgment Boundary in Multi-Agent Systems Security | LLM-based multi-agent systems (MAS) engage tools, share memory, and delegate tasks, often encountering adversarial content. Current de… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Uniform Discrete Diffusion Models are Minimax Optimal for Estimating Distributions with Small Effective Support Size | Discrete diffusion models have emerged as a practically successful framework for generative modeling on discrete product spaces, yet t… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Does On-Policy Distillation for Safety Pose Backdoor Risks? | On-policy distillation (OPD) has attracted growing attention as an effective way to transfer capabilities from teacher models to stude… | ↗Not briefed |