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 | 11d | Personalized Memory Rubrics: Learning What to Remember per User via Rubric-Guided Credit Assignment | Long-horizon personalized agents maintain an evolving memory by continuously deciding what to add, modify, or discard. However, learni… | ↗Not briefed |
arxiv arxiv-ai | 11d | Margin-Guided Token Credit for Reinforcement Learning with Verifiable Rewards | In reinforcement learning with verifiable rewards, a language model is rewarded for a correct final answer, and GRPO passes that rewar… | ↗Not briefed |
bluesky bluesky-ai | 11d | Considering that the AI bubble has finally bursted in China as well as the results of the midterms, I feel that this deal is not going to l… | Considering that the AI bubble has finally bursted in China as well as the results of the midterms, I feel that this deal is not going… | ↗Not briefed |
arxiv arxiv-ai | 11d | A Stage-Wise Explainable Pipeline for Arteriolar-to-Venular Ratio Estimation and Retinopathy Screening | The arteriolar-to-venular diameter ratio (AVR) is a clinically motivated marker of retinal vascular health. We present a five-stage pi… | ↗Not briefed |
arxiv arxiv-ai | 11d | One Sequence Is Enough: Fingerprinting Language Models from Fixed-Length Integer Sequences | Users of model APIs and third-party platforms often rely on the declared model name to identify the model responding to their requests… | ↗Not briefed |
bluesky bluesky-ai | 11d | Househouse screwed Househouse, here. They didn't need to use an LLM. Like band posters and job applications, nobody needed these things un… | Househouse screwed Househouse, here. They didn't need to use an LLM. Like band posters and job applications, nobody needed these thin… | ↗Not briefed |
hn hn-ai | 11d | Show HN: Relinos – Turn Any Website into a REST API (Open-Source) | Hey HN,<p>I made Relinos, an open-source programmable browser that lets you turn websites into reusable APIs.<p>The idea is to make br… | ↗Not briefed |
hn hn-ai | 11d | Anthropic to Seek $1.3T Valuation in IPO | — | ↗Not briefed |
arxiv arxiv-ai | 11d | Coding Agent Memory Post-Training II: Unlocking the Memory Potential of Pre-trained File Operations for Long-Horizon Tasks via On-Policy Distillation | Language-model agents increasingly tackle long-horizon tasks whose interaction histories exceed the model's active context. They must… | ↗Not briefed |
bluesky bluesky-ai | 11d | Das LLM soll nicht einfach aus seinem Trainingswissen Nachrichten zusammenfantasieren. Es bekommt aktuelle Quellen, eine Themenliste und ko… | Das LLM soll nicht einfach aus seinem Trainingswissen Nachrichten zusammenfantasieren. Es bekommt aktuelle Quellen, eine Themenliste u… | ↗Not briefed |
hn hn-ai | 11d | Show HN: Markview – a fast Markdown reader in Rust, Obsidian Vault compatible | Hey HN! I built Markview after coming across RaTeX which is a library renders LaTeX natively without JavaScript or webview. I thought… | ↗Not briefed |
hn hn-ai | 11d | Show HN: AI Vision Index, private tests of how well 36 AI models see | I wanted to know for my work which model is best at vision tasks (using both llms like claude and smaller models like YOLO - smaller s… | ↗Not briefed |
hn hn-ai | 11d | Telling your number parser what to expect | — | ↗Not briefed |
hn hn-ai | 11d | Show HN: A sanctions and PEP screening API | Hey HN, feeling out of my depth posting here. But looking for hard views.<p>Serial regtech founder here and in my last project, roughl… | ↗Not briefed |
arxiv arxiv-ai | 11d | Normalizing Flows for Classical Simulation of Quantum Circuits in Phase Space | We present a novel framework to learn a representation of quantum circuits based on the phase space formulation of many-body systems.… | ↗Not briefed |
arxiv arxiv-ai | 11d | LUMIR: Decoupled Multimodal Representation Learning from Heterogeneous Data for Next-POI Recommendation🔧 Modal | As multimodal tourism-related information collected from heterogeneous data sources has become increasingly available, learning effect… | ↗Not briefed |
bluesky bluesky-ai | 11d | A Sam Altman parody offers an amusing lesson in memory management arstechnica.com/gadgets/2026... | A Sam Altman parody offers an amusing lesson in memory management arstechnica.com/gadgets/2026... | ↗Not briefed |
bluesky bluesky-ai | 11d | LLM-centered development is *thirsting* for legitimacy right now. 🏠🏠 admitting to slipping up is bad - it would be worse if they were giv… | LLM-centered development is *thirsting* for legitimacy right now. 🏠🏠 admitting to slipping up is bad - it would be worse if they wer… | ↗Not briefed |
arxiv arxiv-ai | 11d | Specify Privacy Yourself: Assessing Inference-Time Personalized Privacy Preservation Ability of Large Vision-Language Models | Large Vision-Language Models (LVLMs) have demonstrated remarkable capabilities but raise significant privacy concerns due to their abi… | ↗Not briefed |
bluesky bluesky-ai | 11d | My earnest response to your earnest response will be that, yes, it's totally possible to avoid and it will start by stopping to believe tha… | My earnest response to your earnest response will be that, yes, it's totally possible to avoid and it will start by stopping to believ… | ↗Not briefed |
bluesky bluesky-ai | 11d | seriously though. The team just used an LLM to fix bugs and performance. It made a better game that took less headaches, they shouldn't apo… | seriously though. The team just used an LLM to fix bugs and performance. It made a better game that took less headaches, they shouldn'… | ↗Not briefed |
arxiv arxiv-ai | 11d | BioXArena-Pro: a research-grade benchmark for autonomous biomedical machine learning agents | Coding agents are increasingly proposed as autonomous builders of biomedical models, but existing benchmarks are small, near saturatio… | ↗Not briefed |
bluesky bluesky-ai | 11d | LLMは本質的にダメAIって話、AI業界の人はだれも言わないの、なんで? と思ってたんだけど大御所が言ってた。ヤン・ルカン氏インタビュー。こいつは見ごたえがありますよ。 ・人類滅亡? は? なに言ってんの? ・テキストだけで世界は理解できない ・ワールドモデルが必要 ・体性フ… | LLMは本質的にダメAIって話、AI業界の人はだれも言わないの、なんで? と思ってたんだけど大御所が言ってた。ヤン・ルカン氏インタビュー。こいつは見ごたえがありますよ。 ・人類滅亡? は? なに言ってんの? ・テキストだけで世界は理解できない ・ワールドモデルが必要… | ↗Not briefed |
arxiv arxiv-ai | 11d | Temporal Behavior Trees for Reinforcement Learning: Specifications as Rewards, Curricula, and Diagnostics | Reinforcement learning from formal specifications often separates task definition, reward shaping, curriculum design, and policy diagn… | ↗Not briefed |
bluesky bluesky-ai | 11d | Anthropic: We’re worth $1.4 trillion. Also, Anthropic: We can’t afford to license the copyrighted materials we used to build our models. | Anthropic: We’re worth $1.4 trillion. Also, Anthropic: We can’t afford to license the copyrighted materials we used to build our mode… | ↗Not briefed |
bluesky bluesky-ai | 11d | Your “carbon footprint” is not real. everyone who has ever used an LLM is not individually complicit in the continuation of a data center. … | Your “carbon footprint” is not real. everyone who has ever used an LLM is not individually complicit in the continuation of a data cen… | ↗Not briefed |
hn hn-ai | 11d | Online communities help people w hypermobility disorders connect&get diagnosed | — | ↗Not briefed |
arxiv arxiv-ai | 11d | Pulse or Provenance? What rPPG-Based Deepfake Detectors Learn | Remote photoplethysmography (rPPG) has been proposed as a physiological cue for deepfake detection, based on the premise that synthesi… | ↗Not briefed |
arxiv arxiv-ai | 11d | Universal Approximation Theorem for Metaplectic Neural Operators | We prove universal approximation for metaplectic neural operators on nonempty bounded Lipschitz domains. The nonlocal branches are who… | ↗Not briefed |
hn hn-ai | 11d | Let over Lambda–50 Years of Lisp | — | ↗Not briefed |
hn hn-ai | 11d | Ask HN: Could Stanford's NCM potentially replace modern LLMs in the future? | — | ↗Not briefed |
bluesky bluesky-ai | 11d | People keep providing context to exceptional and/or occasional LLM and generative AI usage as legitimate based on its usefulness. But that'… | People keep providing context to exceptional and/or occasional LLM and generative AI usage as legitimate based on its usefulness. But… | ↗Not briefed |
bluesky bluesky-ai | 11d | That non-aborted kid... now learning to use llm could very well be solving cancer cure in their parent's basement. /s Like seriously agre… | That non-aborted kid... now learning to use llm could very well be solving cancer cure in their parent's basement. /s Like seriously… | ↗Not briefed |
arxiv arxiv-ai | 11d | On the information rate of phones, syllables and words | Which linguistic units minimise the bitrate required to encode speech content: phones, syllables, words, maybe larger units, or someth… | ↗Not briefed |
bluesky bluesky-ai | 11d | updated my review on Big Walk | updated my review on Big Walk | ↗Not briefed |
bluesky bluesky-ai | 11d | Seeing some people defending the use of AI for big walk cuz it's just code and the llm didn't write the code itself and I gotta say, as a p… | Seeing some people defending the use of AI for big walk cuz it's just code and the llm didn't write the code itself and I gotta say, a… | ↗Not briefed |
hn hn-ai | 11d | Show HN: KubeSilicon – Linux VMs and Mac GPU Jobs from Kubernetes Objects | — | ↗Not briefed |
bluesky bluesky-ai | 11d | Breaking News: Anthropic set a price range for its I.P.O. that values the A.I. company at up to $1.4 trillion, a staggering sum. | Breaking News: Anthropic set a price range for its I.P.O. that values the A.I. company at up to $1.4 trillion, a staggering sum. | ↗Not briefed |
bluesky bluesky-ai | 11d | lmao I feel like "a staggering sum" is the closest the nytimes will ever get to saying "you're full of shit" | lmao I feel like "a staggering sum" is the closest the nytimes will ever get to saying "you're full of shit" | ↗Not briefed |
bluesky bluesky-ai | 11d | yknow. im not gonna defend the llm usage but. seeing the energy yall have come up with for this company really highlights the energy you -d… | yknow. im not gonna defend the llm usage but. seeing the energy yall have come up with for this company really highlights the energy y… | ↗Not briefed |
hn hn-ai | 11d | Anthropic Targets Valuation of Up to $1.4T for Blockbuster IPO | — | ↗Not briefed |
hn hn-ai | 11d | Serving open-weight models at scale with vLLM🔧 vLLM | — | ↗Not briefed |
bluesky bluesky-ai | 11d | WaPo columnist further muddies the question of whether AIs are conscious by asking if they have souls. At this rate, we'll soon be debatin… | WaPo columnist further muddies the question of whether AIs are conscious by asking if they have souls. At this rate, we'll soon be de… | ↗Not briefed |
bluesky bluesky-ai | 11d | we do unfortunately still have to [redacted] every anthropic executive and investor. openai too. | we do unfortunately still have to [redacted] every anthropic executive and investor. openai too. | ↗Not briefed |
rss techmeme | 11d | Sources: Microsoft is among a group of companies lobbying the Trump administration to halt the FCC's plan to ban imports of Chinese optical transceivers (Alexandra Alper/Reuters) | Alexandra Alper / Reuters: Sources: Microsoft is among a group of companies lobbying the Trump administration to halt the FCC's plan t… | ↗Not briefed |
hn hn-ai | 11d | Ask HN: What other tools similar to Alibaba OCR do people use with LLMs? | I'm asking the question focusing specifically on what I could use/download/configure right now, not necessarily on any… | ↗Not briefed |
bluesky bluesky-ai | 11d | The way it was described, using an LLM to generate code that was then copy and pasted in and modified to fit, is almost impossible to avoid… | The way it was described, using an LLM to generate code that was then copy and pasted in and modified to fit, is almost impossible to… | ↗Not briefed |
bluesky bluesky-ai | 11d | Anthropic plans to raise up to $64bn in IPO ft.trib.al/gyTJy1L | Anthropic plans to raise up to $64bn in IPO ft.trib.al/gyTJy1L | ↗Not briefed |
bluesky bluesky-ai | 11d | www.thestar.com/news/canada/... | www.thestar.com/news/canada/... | ↗Not briefed |
bluesky bluesky-ai | 11d | おれは各AIによりそいホットラインを固く禁じている MuseとClaudeは禁止すると遵守してくれる | おれは各AIによりそいホットラインを固く禁じている MuseとClaudeは禁止すると遵守してくれる | ↗Not briefed |