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 | MedicalHarness: A Controlled Evaluation of LLMs and Agent Harnesses on Medical Tasks | LLM agents are increasingly built for medical work and scored on clinical benchmarks. Each such score, however, comes from a model run… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Mining Agent Skills from Production Traces | Agent skills that record procedural instructions are increasingly mined from execution traces rather than curated by hand. Skill-minin… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Remember when us phil-tech types said effective altruism was just utilitarianism taken to its worst possible extents? This is what we meant… | Remember when us phil-tech types said effective altruism was just utilitarianism taken to its worst possible extents? This is what we… | ↗Not briefed |
arxiv arxiv-ai | 1mo | AdaSpark: Adaptive DSpark with Online Learning for Tree Verification and N-gram Fill | Block drafters such as DSpark propose ranked candidates for several positions in one forward pass, and a tree verifier checks them in… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Relational Synthesis: Structure-Mediated Concatenative Synthesis for Foley and Retrieval-Augmented Audio Generation | We ask: given a retrieved source audio $S$ and a separate reference audio $R$, can we synthesize novel audio $Y$ out of this pair $(S,… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Meanwhile www.reuters.com/legal/govern... | Meanwhile www.reuters.com/legal/govern... | ↗Not briefed |
bluesky bluesky-ai | 1mo | フロンティアモデルの能力向上に伴ってこうしたペネトレーション報告が激増している…ように見える…んだけど（関係ないという説もありますが、なくはないだろう） 本当に怖いのは、侵入側がLLMを駆使することで、入手した個人情報をさらに大規模かつ高度に、各ユーザーにカスタマイズされた信… | フロンティアモデルの能力向上に伴ってこうしたペネトレーション報告が激増している…ように見える…んだけど（関係ないという説もありますが、なくはないだろう） 本当に怖いのは、侵入側がLLMを駆使することで、入手した個人情報をさらに大規模かつ高度に、各ユーザーにカスタマイ… | ↗Not briefed |
rss marktechpost | 1mo | The Story of Qwen: Alibaba’s AI Models From 7B to 2.4T | Alibaba's Qwen went from an invite-only chatbot in April 2023 to a 2.4-trillion-parameter open-weight model in August 2026. This is th… | ↗Not briefed |
hn hn-ai | 1mo | Why's Foxes: A Guest Book | — | ↗Not briefed |
rss bing-news-ai | 1mo | SpaceX is now a super intelligence company. Say hello to SpaceXSI. | The tech industry has renamed things to fall in line with the Trump administration. | ↗Not briefed |
rss bing-news-ai | 1mo | SpaceX is now a super intelligence company. Say hello to SpaceXSI. | The tech industry has renamed things to fall in line with the Trump administration. | ↗Not briefed |
bluesky bluesky-ai | 1mo | OpenAI CEO Sam Altman has a chilling take on the bargain humans should strike for AI. | OpenAI CEO Sam Altman has a chilling take on the bargain humans should strike for AI. | ↗Not briefed |
hn hn-ai | 1mo | Show HN: Recursant, the agent-aware model router built for hybrid workloads🔧 OpenRouter | Sharing Recursant, the agent-aware model router for private/public workloads. Cut your AI agent bill by 30% and seamlessly operat… | ↗Not briefed |
hn hn-ai | 1mo | Show HN: SRA-RiskGate-4B – Local LLM for stablecoin risk gating and disputes | — | ↗Not briefed |
arxiv arxiv-ai | 1mo | Beyond Semantic Similarity: Performance and Costs of Agentic Retrieval for Complex Tasks | Modern information systems, including many agentic workflows, use dense retrieval to explore large amounts of unstructured data. Howev… | ↗Not briefed |
bluesky bluesky-ai | 1mo | office door closed, someone knocks. i quickly alt-tab over to the lewd pics, lest anyone think i'm reading LLM papers | office door closed, someone knocks. i quickly alt-tab over to the lewd pics, lest anyone think i'm reading LLM papers | ↗Not briefed |
bluesky bluesky-ai | 1mo | get you an LLM that can make you believe crazy things and then talk you down after | get you an LLM that can make you believe crazy things and then talk you down after | ↗Not briefed |
arxiv arxiv-ai | 1mo | Beyond In-Distribution Preservation: Recovering Generalization in Quantized VLAs via Vulnerability-Oriented Tuning | Post-training quantization has been shown to preserve VLA performance under standard evaluation conditions, but whether it preserves t… | ↗Not briefed |
arxiv arxiv-ai | 1mo | CIPHER-MoE: Balancing Efficiency and Routing Fidelity in Trillion-Scale MoE Training | Mixture-of-Experts (MoE) has been widely adopted in recent large language model (LLM) architectures. However, scaling up MoE in LLM tr… | ↗Not briefed |
bluesky bluesky-ai | 1mo | We’re using a custom BirdNET model to automate ID of 140 bird species is south-east Australia. This relies on a convolutional neural networ… | We’re using a custom BirdNET model to automate ID of 140 bird species is south-east Australia. This relies on a convolutional neural n… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Trump names national intelligence director Jay Clayton to lead new federal AI task force #ausnews #altman #anthropic #OpenAI #ChatGPT #nvid…🔧 ChatGPT | Trump names national intelligence director Jay Clayton to lead new federal AI task force #ausnews #altman #anthropic #OpenAI #ChatGPT… | ↗Not briefed |
bluesky bluesky-ai | 1mo | @iruletheworldmo.bsky.social forget astra 6.1 something much bigger is coming #AI #Anthropic #OpenAI #SSI ( Donnie is terrible at this … | @iruletheworldmo.bsky.social forget astra 6.1 something much bigger is coming #AI #Anthropic #OpenAI #SSI ( Donnie is terrible at… | ↗Not briefed |
arxiv arxiv-ai | 1mo | HLA-WM: Hybrid Linear Attention for Long-Horizon Video World Models | Long-horizon video world models require persistent memory to preserve scene consistency over extended rollouts. Softmax attention reta… | ↗Not briefed |
bluesky bluesky-ai | 1mo | If you actually read the article, he is actually being tame and agreeable here, talking about allowing e.g open models instead of gatekeepi… | If you actually read the article, he is actually being tame and agreeable here, talking about allowing e.g open models instead of gate… | ↗Not briefed |
hn hn-ai | 1mo | The AI Safety community is unfortunately doing more harm than good | — | ↗Not briefed |
arxiv arxiv-ai | 1mo | PACMI: Provenance-Aware Cascading Memory Invalidation for Long-Term LLM Agents | LLM agents rely on long-term memory to retain and reuse information when performing tasks over long horizons. Existing methods provide… | ↗Not briefed |
rss techmeme | 1mo | Sources: several Western open-weight models are set to launch this month, including Reflection AI's first model, which will rival top Chinese open-weight models (Bradley Olson/Axios) | Bradley Olson / Axios: Sources: several Western open-weight models are set to launch this month, including Reflection AI's first model… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Like this is as plainly and neutral something this wide-appealing and used can get before stoking flames really | Like this is as plainly and neutral something this wide-appealing and used can get before stoking flames really | ↗Not briefed |
arxiv arxiv-ai | 1mo | FreSia: Frequency-Semantic Instantiation and Alignment for Multivariate Time Series Analysis🔧 Modal | Large Language Models (LLMs) have shown strong potential in multivariate time series forecasting and anomaly detection. Existing studi… | ↗Not briefed |
hn hn-ai | 1mo | Show HN: Ramen – self-hosted multi-zone MCP server for Kubernetes(Rust & Python) | — | ↗Not briefed |
bluesky bluesky-ai | 1mo | llm trained on user was blocked for this post | llm trained on user was blocked for this post | ↗Not briefed |
bluesky bluesky-ai | 1mo | 🚨 AI NEWS: Sam Altman, the owner of OpenAI, molested his sister for 9 years and this is just the first time you are likely hearing about i… | 🚨 AI NEWS: Sam Altman, the owner of OpenAI, molested his sister for 9 years and this is just the first time you are likely hearing ab… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Sam Altman says "the world should accept some bad things happening for the benefit of AI." Perhaps Altman being charged, tried and convicte… | Sam Altman says "the world should accept some bad things happening for the benefit of AI." Perhaps Altman being charged, tried and con… | ↗Not briefed |
bluesky bluesky-ai | 1mo | his sister has repeatedly accused him of sexual abuse www.reuters.com/legal/govern... | his sister has repeatedly accused him of sexual abuse www.reuters.com/legal/govern... | ↗Not briefed |
arxiv arxiv-ai | 1mo | Inferring physical fields in coupled systems with unknown parameters from incomplete observations using physics-constrained attentive neural operators | Given incomplete measurements of a single physical field in a coupled system with unknown parameters, can we infer its full physical s… | ↗Not briefed |
bluesky bluesky-ai | 1mo | I just keep thining about, the zillions of dollars in fake money that are being generated in SpaceX & Anthropic stock. How much of that is … | I just keep thining about, the zillions of dollars in fake money that are being generated in SpaceX & Anthropic stock. How much of tha… | ↗Not briefed |
bluesky bluesky-ai | 1mo | this slow ass deepseek provider will fix a bug in around the time it takes sol to create a civilization and then make every citizen wear co… | this slow ass deepseek provider will fix a bug in around the time it takes sol to create a civilization and then make every citizen we… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Yeah, I did see that! I had a private one too recently where I was like time to hit the bricks because nobody really read anything I said a… | Yeah, I did see that! I had a private one too recently where I was like time to hit the bricks because nobody really read anything I s… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Why does OpenAI need a CEO anymore if its AI is so good? Why not replace your CEO with an AI? | Why does OpenAI need a CEO anymore if its AI is so good? Why not replace your CEO with an AI? | ↗Not briefed |
hn hn-ai | 1mo | Ask HN: What did you think about Digger (2026)? | I just got back from the cinemas and cannot stop thinking about the takes it had on where we are today…especially the ode to AI lol. N… | ↗Not briefed |
arxiv arxiv-ai | 1mo | ReMaD: Tuning-free Domain Adaptation for Classification and Out-of-Distribution Detection | We introduce Reduced-rank Mahalanobis Distance (ReMaD), a novel prototypical distance-based refinement to classification and out-of-di… | ↗Not briefed |
bluesky bluesky-ai | 1mo | He should be in an Australian prison for OpenAI hacking their government. If you or I did it, that’s where we’d be. We need to start lockin… | He should be in an Australian prison for OpenAI hacking their government. If you or I did it, that’s where we’d be. We need to start l… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Retrieval-Based In-Context Learning: A Domain Adaptation Framework | In-context retrieval (ICR) is a retrieval-based form of in-context learning (ICL) in which demonstrations are retrieved from a source… | ↗Not briefed |
bluesky bluesky-ai | 1mo | normal product liability law is pretty foreign to consumer internet companies… …but, this line of talk from Altman may sink OpenAI overall… | normal product liability law is pretty foreign to consumer internet companies… …but, this line of talk from Altman may sink OpenAI ov… | ↗Not briefed |
bluesky bluesky-ai | 1mo | It is a compilation of both Australian, US and international news regarding AI, hacking and theft. We decided to put all the reporting on … | It is a compilation of both Australian, US and international news regarding AI, hacking and theft. We decided to put all the reportin… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Close, but reversed: Altman should accept some bad things happening to his OpenAI for the sake of our world. | Close, but reversed: Altman should accept some bad things happening to his OpenAI for the sake of our world. | ↗Not briefed |
arxiv arxiv-ai | 1mo | Rotated, but How Far? Diagnosing and Improving Object-Rotation Reasoning in VLMs | Vision-language models (VLMs) can detect that an object has rotated across views, but cannot reliably tell by how much. We introduce O… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Altman said unlike rival Anthropic, his company believes in a “lighter-touch regulatory stance” on AI and that it will be used for more “go… | Altman said unlike rival Anthropic, his company believes in a “lighter-touch regulatory stance” on AI and that it will be used for mor… | ↗Not briefed |
rss techmeme | 1mo | A look at Sean Parker's resurrection of Stability AI following Emad Mostaque's ousting, and its new focus on AI for music professionals, backed by major labels (Abram Brown/The Information) | Abram Brown / The Information: A look at Sean Parker's resurrection of Stability AI following Emad Mostaque's ousting, and its new foc… | ↗Not briefed |
rss bing-news-ai | 1mo | Donald Trump’s New ‘Super Intelligence Force’ Has an Unexpected Leader | President Donald Trump surprised Washington on Sunday by appointing Jay Clayton, the Director of National Intelligence, to head a newl… | ↗Not briefed |