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 | SciExam for ENSO: Can AI Agents Build Climate Models? | Language-model agents are increasingly asked to carry out open-ended scientific research, yet their results are usually graded against… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Your Prompt Should Do More: Effects of Retrieval Instructions in Embedding Models | Prompted embedding models have recently received increasing attention, particularly for retrieval, where detailed retrieval instructio… | ↗Not briefed |
bluesky bluesky-ai | 1mo | like for one example I am not surprised the openai stuff has lots of weird variables, this is what happens on long horizon projects: it inv… | like for one example I am not surprised the openai stuff has lots of weird variables, this is what happens on long horizon projects: i… | ↗Not briefed |
arxiv arxiv-ai | 1mo | RECAST: Learning to Compute the Right Context through Adaptive Evidence Routing | Large language models are increasingly applied to tasks grounded in long, heterogeneous information sources. Conventional Retrieval-Au… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Validity Without Ground Truth: What Stated-Preference Economics Offers the Evaluation of Language Models | Many of the questions now put to large language models have no correct answer to score against: what a policy is worth, which option a… | ↗Not briefed |
hn hn-ai | 1mo | No new Ghostty updates since March | — | ↗Not briefed |
bluesky bluesky-ai | 1mo | 🔥 The first episode from Offensive AI Con is live! IBM X-Force’s Valentina Palmiotti (chompie) walks us through the autonomous Windows k… | 🔥 The first episode from Offensive AI Con is live! IBM X-Force’s Valentina Palmiotti (chompie) walks us through the autonomous Wind… | ↗Not briefed |
rss the-decoder | 1mo | Google says 180 billion images and videos now carry SynthID watermarks as detector goes public | Google's AI watermark detector SynthID is now public. Anyone can check whether images, videos, or audio were created by Google's AI or… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Oracle-Efficient and Parameter-Free Agnostic Smoothed Online Learning | Online learning is an attractive framework in many domains because it permits well-defined learning even when data are dependent or ch… | ↗Not briefed |
arxiv arxiv-ai | 1mo | EmbodiedRSI: Active Continual Robot Learning Through Hypothesis-Guided Co-Evolution | Robot foundation models provide strong visuomotor control, yet their performance can degrade when object positions or task instruction… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Evolutionary Architecture Search for Chlorophyll-$a$ Prediction in Lakes using Sentinel-2 | Small tabular datasets with expert-designed spectral features are the norm in operational Earth observation, and the networks applied… | ↗Not briefed |
rss theverge-ai | 1mo | Surface RTX Spark Dev Box is available for preorder for $5,999 | Microsoft's Nvidia-powered Surface RTX Spark Dev Box is available for preorder now directly, and slated to ship in November for just a… | ↗Not briefed |
bluesky bluesky-ai | 1mo | מצד שני אם המודלים של openai יכתבו ערעורים כמו שהם כותבים אותם לסודעטים יהיה שמייח | מצד שני אם המודלים של openai יכתבו ערעורים כמו שהם כותבים אותם לסודעטים יהיה שמייח | ↗Not briefed |
arxiv arxiv-ai | 1mo | Best Arm Identification for Bandits with Shifting Means | We study the best arm identification problem in a stochastic environment with a novel form of adversarial perturbations, which we coin… | ↗Not briefed |
bluesky bluesky-ai | 1mo | trying to tell it what i neeeeeed its not sophisticated enough to understand i am interested in gunk from poland but also gunk brand polis… | trying to tell it what i neeeeeed its not sophisticated enough to understand i am interested in gunk from poland but also gunk brand… | ↗Not briefed |
hn hn-ai | 1mo | DuckDB Ducklake | — | ↗Not briefed |
bluesky bluesky-ai | 1mo | Is nothing sacred? | Is nothing sacred? | ↗Not briefed |
arxiv arxiv-ai | 1mo | Two-Level Softmax Sampling Done Right: Correcting Bias from Size Imbalance and Dispersion | Sampling from a softmax distribution is a fundamental operation in machine learning, but its linear complexity in the number of items… | ↗Not briefed |
hn hn-ai | 1mo | Show HN - QRCode URLs scan API | Real-time URL safety API for QR scanners and AI agents. One API call before openURL, or before your agent fetches the payload. | ↗Not briefed |
hn hn-ai | 1mo | Show HN: ChillSweep – A better cleaner app for PC and Mac | I was frustrated with the manual busy work of cleaning appdata leftovers and how much most traditional cleaner apps miss. Most cleaner… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Before They Can Solve: Predicting Post-Training Coding-Agent Performance from Base Models | How can we predict which base checkpoint is worth an expensive round of agentic post-training? End-to-end pass@$K$ tests whether succe… | ↗Not briefed |
hn hn-ai | 1mo | Digital Transformation ROI in 2026 | — | ↗Not briefed |
bluesky bluesky-ai | 1mo | Yep, you are using that term correctly. In AI-speak I usually see the term “router” (where it routes to the best model) but wrapper is simi… | Yep, you are using that term correctly. In AI-speak I usually see the term “router” (where it routes to the best model) but wrapper is… | ↗Not briefed |
bluesky bluesky-ai | 1mo | 🤖 Beyond the Ground · AI & DeepTech Toss Bank's AI Counseling and the Arithmetic of '300 Hours Saved': The 5x Paradox of Suggestion Chips… | 🤖 Beyond the Ground · AI & DeepTech Toss Bank's AI Counseling and the Arithmetic of '300 Hours Saved': The 5x Paradox of Suggestion… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Anyone pretending LLM watermarking is effective at much of anything is trying to make money or is absolutely clueless. There’s no third thi… | Anyone pretending LLM watermarking is effective at much of anything is trying to make money or is absolutely clueless. There’s no thir… | ↗Not briefed |
hn hn-ai | 1mo | Driving an Nvidia GPU from a Hobby Rust Kernel: GSP-RM to Vulkan | — | ↗Not briefed |
arxiv arxiv-ai | 1mo | A Society of Researchers: Designing Institutions for Populations of Autonomous Research Agents | Deployments of research agents are moving to populations of thousands that share one pool of compute, while most current systems organ… | ↗Not briefed |
bluesky bluesky-ai | 1mo | "Inspired by his call to use AI, I went ahead and asked Claude (Anthropic’s AI assistant) to analyze some of Kennedy’s more memorable state… | "Inspired by his call to use AI, I went ahead and asked Claude (Anthropic’s AI assistant) to analyze some of Kennedy’s more memorable… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Ich kann mir keinen Grund denken, warum Anthropic und Co Horrorstories über ihr eigenes Produkt erzählen, in denen das Produkt immer, immer… | Ich kann mir keinen Grund denken, warum Anthropic und Co Horrorstories über ihr eigenes Produkt erzählen, in denen das Produkt immer,… | ↗Not briefed |
bluesky bluesky-ai | 1mo | In @nytopinion.nytimes.com “I think OpenAI and its peers are now producing a technology that is more capable and poses more risk than wha… | In @nytopinion.nytimes.com “I think OpenAI and its peers are now producing a technology that is more capable and poses more risk tha… | ↗Not briefed |
arxiv arxiv-ai | 1mo | How assigned AI use before class shapes active student engagement in class | AI learning tools are rapidly entering classrooms, but evidence about whether they help students learn is mixed and rests mostly on te… | ↗Not briefed |
arxiv arxiv-ai | 1mo | FoldBack: Self-Correcting Masked Generative Policy for Long-Horizon Garment Folding | We present FoldBack, a self-correcting masked generative policy for long-horizon garment folding. Existing long-trajectory policies ma… | ↗Not briefed |
hn hn-ai | 1mo | You click a PostHog survey. An attacker gets your whole account | — | ↗Not briefed |
arxiv arxiv-ai | 1mo | Composing What Each Teacher Learned: Multi-Teacher On-Policy Distillation through Teacher-Relative Shifts | Multi-teacher on-policy distillation (MOPD) is used in two settings. In common-domain composition, several teachers score each student… | ↗Not briefed |
arxiv arxiv-ai | 1mo | NeuralBES: A Differentiable, Control-Aware Emulator for Scalable Building Energy Modeling | Demand-side flexibility i.e. forecasting, shifting, and curtailing residential energy loads, depends on thermal models trusted across… | ↗Not briefed |
bluesky bluesky-ai | 1mo | this is the future of math. every once in a while openAI drops a mountain of crap and says “you losers go figure it out.” and if you’re an … | this is the future of math. every once in a while openAI drops a mountain of crap and says “you losers go figure it out.” and if you’r… | ↗Not briefed |
hn hn-ai | 1mo | Show HN: Manifestack – open-source stack audit inside your coding agent | — | ↗Not briefed |
arxiv arxiv-ai | 1mo | PHRBench: A Behavioral Evaluation of Post-Hallucination Reasoning in LLMs | Hallucinated information can propagate through multi-stage LLM systems and become part of the context for subsequent reasoning. Existi… | ↗Not briefed |
hn hn-ai | 1mo | I found an explanet using Claude, and now you can fly to it yourself | — | ↗Not briefed |
bluesky bluesky-ai | 1mo | youtube.com/watch?v=vhWe... | youtube.com/watch?v=vhWe... | ↗Not briefed |
bluesky bluesky-ai | 1mo | Microsoft Introduces AI Coding Model That Runs on Laptops; DeepSeek V4 Flash Runs on 60GB Machines, Outperforming GPT-5 on Some Coding Task…🔧 GPT-5 | Microsoft Introduces AI Coding Model That Runs on Laptops; DeepSeek V4 Flash Runs on 60GB Machines, Outperforming GPT-5 on Some Coding… | ↗Not briefed |
hn hn-ai | 1mo | Github Status - Incident with Git Operations, Issues, Actions and Pull Requests | — | ↗Not briefed |
rss techmeme | 1mo | Meta rolls out AI tools to find ads that secretly lead users to CSAM and says it took action against 33.2M pieces of child exploitation content in H1 2026 (Lauren Forristal/TechCrunch) | Lauren Forristal / TechCrunch: Meta rolls out AI tools to find ads that secretly lead users to CSAM and says it took action against 33… | ↗Not briefed |
arxiv arxiv-ai | 1mo | A Good Self-Teacher Meets the Student Where They Are: Joint On-Policy Learning and Teaching | Reinforcement Learning (RL) from outcome rewards suffers from sparse supervision, particularly on difficult, long-horizon tasks where… | ↗Not briefed |
bluesky bluesky-ai | 1mo | On the one hand it's not much, on the other hand I also think this is kind of a big deal. Apple at least works with Common Sense and includ… | On the one hand it's not much, on the other hand I also think this is kind of a big deal. Apple at least works with Common Sense and i… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Preeti and Cassie, avoid them!!! (They tracked me down via means they should have had no access to.) authorsguild.org/news/claimsh... | Preeti and Cassie, avoid them!!! (They tracked me down via means they should have had no access to.) authorsguild.org/news/claimsh... | ↗Not briefed |
bluesky bluesky-ai | 1mo | unfortunately i cant come up with a more sound model of this because i cant stomach looking at the twitter side of things anymore, i almost… | unfortunately i cant come up with a more sound model of this because i cant stomach looking at the twitter side of things anymore, i a… | ↗Not briefed |
hn hn-ai | 1mo | A thorium-229 optical nuclear clock with feedback loop | — | ↗Not briefed |
arxiv arxiv-ai | 1mo | RunningTab: Direct Workspace Interaction with Environment-Side Tabs | Much knowledge work produces new deliverables from files a workspace already holds, and LLM agents are beginning to take such work ove… | ↗Not briefed |
hn hn-ai | 1mo | Show HN: Terse, a Claude Code plugin that halves reply length by cutting filler | I struggle with the way Claude speaks to me when I do work. So much text, every, single, time.<p>When I ask a question, I get an essay… | ↗Not briefed |