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 | A Composable AI-Accelerated Iterative Solver for 3D-IC Thermal Modeling🔧 FLUX | Accurate thermal analysis of heterogeneous 2.5D/3D-IC packages is essential yet computationally prohibitive. A single full-package FEM… | ↗Not briefed |
arxiv arxiv-ai | 1mo | CUEing User Simulators: Calibrated User Embeddings for Multi-Turn Benchmarking | Recent benchmarks rely on user simulators to evaluate AI agents in multi-turn interaction. While existing simulation techniques demons… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Solid chance Openai goes belly up and takes softbank and oracle with it. Couldn't happen to more deserving companies. | Solid chance Openai goes belly up and takes softbank and oracle with it. Couldn't happen to more deserving companies. | ↗Not briefed |
bluesky bluesky-ai | 1mo | Mayo Clinic AI spots pancreatic cancer early A Mayo Clinic AI model identified high-risk individuals for pancreatic cancer up to three yea… | Mayo Clinic AI spots pancreatic cancer early A Mayo Clinic AI model identified high-risk individuals for pancreatic cancer up to thre… | ↗Not briefed |
hn hn-ai | 1mo | Show HN: Rhun, an open-source code editor written in assembly🔧 Ollama | I found that I'm not using even 1/3 of vim/vscode features anymore.<p>That's wht I'm building rhun - a small… | ↗Not briefed |
hn hn-ai | 1mo | E2E: Next generation E2E testing framework for web and mobile apps | — | ↗Not briefed |
hn hn-ai | 1mo | Show HN: HeadWater Volume Ingestion Matrix – a production shortcut | — | ↗Not briefed |
arxiv arxiv-ai | 1mo | SideKernel: A Usable microVM Sandbox for AI Coding Agents on macOS | AI coding agents are untrusted system components, yet they require autonomy on the developer machines they run on. This contradiction… | ↗Not briefed |
arxiv arxiv-ai | 1mo | FinDialogLens: Event Extraction over Multi-Party Dialogue for Missed-Trade Identification in Financial Chatrooms | Multi-party financial chatrooms are vital for sales-and-trading professionals, but their complexity makes manual recovery of missed tr… | ↗Not briefed |
bluesky bluesky-ai | 1mo | This, but replace "...AI Agents" with "...Hype Videos" www.nature.com/articles/s41... | This, but replace "...AI Agents" with "...Hype Videos" www.nature.com/articles/s41... | ↗Not briefed |
rss arstechnica-tech | 1mo | Hacks of 2 federal agencies in a month have spilled a bonanza of sensitive data | It has been a bad month for federal government cybersecurity. | ↗Not briefed |
hn hn-ai | 1mo | Nvidia debuted the Open Agent Safety Platform | — | ↗Not briefed |
bluesky bluesky-ai | 1mo | Still seething that my ex-publisher didn't bother to copyright my 4 novels and even though I took back all the rights years ago and they sh… | Still seething that my ex-publisher didn't bother to copyright my 4 novels and even though I took back all the rights years ago and th… | ↗Not briefed |
bluesky bluesky-ai | 1mo | the furry antipathy towards AI art is actually a memetic defence against the cognitohazard that “getting an image model to edit all your IR… | the furry antipathy towards AI art is actually a memetic defence against the cognitohazard that “getting an image model to edit all yo… | ↗Not briefed |
hn hn-ai | 1mo | Manyfold's Agents.md Is Based | — | ↗Not briefed |
arxiv arxiv-ai | 1mo | Reinforcement Learning Techniques for the Optimization of Target Polarization in Nuclear Physics Scattering Experiments | The operation of dynamically polarized targets in nuclear physics experiments relies on continuous tuning of the microwave frequency t… | ↗Not briefed |
bluesky bluesky-ai | 1mo | AI models can't agree on what humans think 🤖 Pew Research Center found that synthetic polling by GPT-5.1 and Claude Opus 4.6 produced inc…🔧 GPT-5 | AI models can't agree on what humans think 🤖 Pew Research Center found that synthetic polling by GPT-5.1 and Claude Opus 4.6 produce… | ↗Not briefed |
bluesky bluesky-ai | 1mo | random thought but LLM slop is gonna be terrible for that entire sub-industry of making silly budget games that aren't good and are a lot o… | random thought but LLM slop is gonna be terrible for that entire sub-industry of making silly budget games that aren't good and are a… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Josh Hawley exists exclusively to help Republicans pretend they support meaningful corporate oversight | Josh Hawley exists exclusively to help Republicans pretend they support meaningful corporate oversight | ↗Not briefed |
bluesky bluesky-ai | 1mo | honestly ffs people make insanely good short game jams games all the time that leave an impact. like every time i think "oh my god am i bei… | honestly ffs people make insanely good short game jams games all the time that leave an impact. like every time i think "oh my god am… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Religious Scholars Met With Anthropic. What They Heard Stunned Them. | NY Times 🎁 article www.nytimes.com/2026/09/29/u... | Religious Scholars Met With Anthropic. What They Heard Stunned Them. | NY Times 🎁 article www.nytimes.com/2026/09/29/u... | ↗Not briefed |
bluesky bluesky-ai | 1mo | everyone's going to avoid it anyway because of the llm usage that's ruined our futures. we can wait the 30 years for it to be done and we'l… | everyone's going to avoid it anyway because of the llm usage that's ruined our futures. we can wait the 30 years for it to be done and… | ↗Not briefed |
bluesky bluesky-ai | 1mo | I’m planning some courses on agentic AI for physics researchers. I have in mind three independent 4-week modules using Claude Code or Code… | I’m planning some courses on agentic AI for physics researchers. I have in mind three independent 4-week modules using Claude Code or… | ↗Not briefed |
bluesky bluesky-ai | 1mo | That's... not what we were expecting. | That's... not what we were expecting. | ↗Not briefed |
bluesky bluesky-ai | 1mo | “…virtual try-on allows ChatGPT users to upload a selfie or a full-body photo in order to visualize how an article of clothing or an access…🔧 ChatGPT | “…virtual try-on allows ChatGPT users to upload a selfie or a full-body photo in order to visualize how an article of clothing or an a… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Introducing LLM Bluesky Golf 🦋⛳ Pick a post. Try to recreate it with an LLM using the shortest, least-specific prompt you can. The fun i… | Introducing LLM Bluesky Golf 🦋⛳ Pick a post. Try to recreate it with an LLM using the shortest, least-specific prompt you can. The… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Counterexample Generation via Per-Theorem Symbolic Verifiers: When Imitation Hurts and Reinforcement Repairs | Large language models often solve a theorem forward yet fail to disprove a closely related false one: a falsification gap that supervi… | ↗Not briefed |
hn hn-ai | 1mo | ArXiv's Updated Rate Limit Policy | — | ↗Not briefed |
hn hn-ai | 1mo | The failure modes of Claude Code in a guided 60h project | — | ↗Not briefed |
hn hn-ai | 1mo | Show HN: HeadWater B2B Data and AI Kit – A Production Architecture Shortcut | — | ↗Not briefed |
arxiv arxiv-ai | 1mo | AI-driven Thermal-aware Data Center Capacity Planning | The emerging of large language models (LLMs) has posed significant challenges to the thermal management of data center. Intense GPU co… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Les mots justes www.liberation.fr/idees-et-deb... | Les mots justes www.liberation.fr/idees-et-deb... | ↗Not briefed |
bluesky bluesky-ai | 1mo | deepseek caiu, impossivel trabalhar | deepseek caiu, impossivel trabalhar | ↗Not briefed |
bluesky bluesky-ai | 1mo | A bizarre epistemics where an LLM having been involved at all makes someone saying "I (a human) have read all this and tested it and it doe… | A bizarre epistemics where an LLM having been involved at all makes someone saying "I (a human) have read all this and tested it and i… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Bandits via Additive Quantized Representations | Contextual bandits require balancing nonlinear reward modeling with online efficiency. Tree ensembles and neural methods capture nonli… | ↗Not briefed |
arxiv arxiv-ai | 1mo | A Generative Model of Complex Networks Using Graphons and Neural Inverse Operators | Generative graph models are central to understanding and simulating complex networks. However, existing approaches have complementary… | ↗Not briefed |
hn hn-ai | 1mo | How the Bayh-Dole Act Transformed Tech Transfer | — | ↗Not briefed |
arxiv arxiv-ai | 1mo | Are you Synthesizing or Recalling? Evaluating LLMs on Algorithmic Code Retrieval | Large language models (LLMs) have demonstrated strong performance in code generation, where success depends on both recalling relevant… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Learning Style, Forgetting Semantics: A Case Study of SFT and RFT on Classification Tasks | Why does supervised fine-tuning (SFT) lead to more forgetting than reinforcement fine-tuning (RFT), even when all teacher demonstratio… | ↗Not briefed |
bluesky bluesky-ai | 1mo | now, because llm codegen, and for no other reason, definitely not because it's not good or i'd never promote it, this won't catch on so an… | now, because llm codegen, and for no other reason, definitely not because it's not good or i'd never promote it, this won't catch on… | ↗Not briefed |
rss techmeme | 1mo | DoorDash pulls support for a GOP bill that would have limited DC's ability to write its own tax laws after widespread calls for locals to boycott the service (Martin Austermuhle/The Washington Sun) | Martin Austermuhle / The Washington Sun: DoorDash pulls support for a GOP bill that would have limited DC's ability to write its own t… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Worth note: at the time of this atrocity Anthropic was the LLM powering the Maven Smart System in partnership with Palantir. Don't let Da… | Worth note: at the time of this atrocity Anthropic was the LLM powering the Maven Smart System in partnership with Palantir. Don't l… | ↗Not briefed |
bluesky bluesky-ai | 1mo | "Anthropic berichtete kürzlich, das KI-Modell Claude erledige 26% der hauseigenen KI-Forschungs- und Entwicklungsarbeiten bereits weitgehen… | "Anthropic berichtete kürzlich, das KI-Modell Claude erledige 26% der hauseigenen KI-Forschungs- und Entwicklungsarbeiten bereits weit… | ↗Not briefed |
rss wired-ai | 1mo | Trump’s ‘Morally Binding’ AI ‘Accord,’ the Rise of AI Agents, and Extremists on the Ballot | This week on “Uncanny Valley,” we discuss the voluntary AI safety agreement tech executives signed, AI agents for normies, and extremi… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Evaluating and Improving the Robustness of Large Language Models to Input Sequence Variations | Large language models (LLMs) in production systems face prompt injections, trojans (backdoors), and manipulation of automatic quality… | ↗Not briefed |
arxiv arxiv-ai | 1mo | CRISP: A Framework for Clause-Reconstructed Interpretable NeuroSymbolic Propositions | Deep neural networks achieve high accuracy through layered numerical transformations, yet their decisions remain difficult to audit be… | ↗Not briefed |
bluesky bluesky-ai | 1mo | ChatGPT wants to be your virtual fitting room OpenAI is rolling out virtual try-ons for clothes and accessories using your photo, plus a F…🔧 ChatGPT | ChatGPT wants to be your virtual fitting room OpenAI is rolling out virtual try-ons for clothes and accessories using your photo, plu… | ↗Not briefed |
hn hn-ai | 1mo | Show HN: MoA as Universal IL –> 30yrs from PhD to 5-device validated compiler | — | ↗Not briefed |
hn hn-ai | 1mo | Free hosted API for Laya, the open-weight decision model | — | ↗Not briefed |
bluesky bluesky-ai | 1mo | ChatGPT developer OpenAI has fired three employees who shared confidential information with tech safety researchers, the Wall Street Journa…🔧 ChatGPT | ChatGPT developer OpenAI has fired three employees who shared confidential information with tech safety researchers, the Wall Street J… | ↗Not briefed |