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 |
|---|---|---|---|---|
hn hn-ai | 1mo | Show HN: OpenPhoto – Photoshop Rebuilt in Rust | — | ↗Not briefed |
arxiv arxiv-ai | 1mo | The Attribution Blind Spot: Layerwise Trajectory Diagnostics for Source Reliance in Retrieval-Augmented Language Models | A retrieval-augmented model can match a document without relying on it. Controlled knowledge conflicts make source choice observable a… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Goldsmith: Gold-Loss-Guided Definition Optimization with an Agentic Annotation Harness | Many annotation projects begin before experts have a stable guideline or enough labels to train a task-specific model. We present Gold… | ↗Not briefed |
bluesky bluesky-ai | 1mo | OpenAI quietly releases the full proofs of 377 open conjectures in various fields of mathematics, deepening maths existential crisis. githu… | OpenAI quietly releases the full proofs of 377 open conjectures in various fields of mathematics, deepening maths existential crisis.… | ↗Not briefed |
bluesky bluesky-ai | 1mo | He's like if an LLM was technically a person. | He's like if an LLM was technically a person. | ↗Not briefed |
arxiv arxiv-ai | 1mo | Correspondences as Decisions: JevNexus for Decision-Centric Schema Matching | Schema matching increasingly uses generative language models to rerank retrieved column candidates, although the underlying task is a… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Mitigating Accent-Language Confusion in Self-Supervised Speech Representations for Language Identification | Spoken language identification (LID) aims to recognize the target language regardless of accent. In practice, however, LID models fine… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Never let an LLM speak for you! | Never let an LLM speak for you! | ↗Not briefed |
hn hn-ai | 1mo | Shaders, WebGPU Components for React, Vue, Svelte, Solid, JavaScript and Framer | — | ↗Not briefed |
arxiv arxiv-ai | 1mo | MIMESIS: Learning User Simulators as Training Environments for Interactive Agents🔧 GPT-5 | Training and evaluating interactive language agents typically requires rich user interactions, yet collecting human feedback is expens… | ↗Not briefed |
bluesky bluesky-ai | 1mo | gotta be honest i really don't get the appeal of throwing an llm at making mods n stuff for you | gotta be honest i really don't get the appeal of throwing an llm at making mods n stuff for you | ↗Not briefed |
rss techmeme | 1mo | Sources: Singapore distances itself from regulating Hyperliquid, citing its "decentralized nature", even as the crypto futures DEX confirms it is based there (Owen Walker/Financial Times) | Owen Walker / Financial Times: Sources: Singapore distances itself from regulating Hyperliquid, citing its “decentralized nature”, eve… | ↗Not briefed |
arxiv arxiv-ai | 1mo | CHASE: Channel-Aligned Structure Exploitation for Geometry-Aware Model Engineering | Geometric and Spectral Alignment (GSA) characterizes trained networks through spectral concentration, physical-channel alignment, supp… | ↗Not briefed |
hn hn-ai | 1mo | Chess Openings by Nationality | — | ↗Not briefed |
arxiv arxiv-ai | 1mo | MORA: Modeling Observed Changes for Drift-Robust Time-Series Anomaly Detection | Time-series anomaly detection (TSAD) identifies deviations from patterns learned from historical data. In non-stationary settings, dis… | ↗Not briefed |
hn hn-ai | 1mo | Show HN: Musql, JIT-compiled SQL, up to 300x faster than SQLite, 1000x vs. Turso | Howdy HN!<p>Missing a fast pure Go embedded SQL database. The issue is that compared to C, go is slow. Having done a lot of JIT in go,… | ↗Not briefed |
bluesky bluesky-ai | 1mo | i feel like it would be hard to tell if mr beast wrote a book or if he got ai to do it, because hes a lot like an llm in many ways | i feel like it would be hard to tell if mr beast wrote a book or if he got ai to do it, because hes a lot like an llm in many ways | ↗Not briefed |
hn hn-ai | 1mo | The Alchemical Transformations of the Mutus Liber (1677) | — | ↗Not briefed |
bluesky bluesky-ai | 1mo | it's also a transparent attempt to juice valuation by way of flooding the space with unreviewed "proofs", occluding that this represents $$… | it's also a transparent attempt to juice valuation by way of flooding the space with unreviewed "proofs", occluding that this represen… | ↗Not briefed |
bluesky bluesky-ai | 1mo | This is the weirdest thing I've seen in quite a while. The world is going crazy. Claude 5.5 made a video attacking the idea of AI. It's inc… | This is the weirdest thing I've seen in quite a while. The world is going crazy. Claude 5.5 made a video attacking the idea of AI. It'… | ↗Not briefed |
bluesky bluesky-ai | 1mo | (Also, Tennessee’s record keeping and their two hundred years of using biological LLM confabulation means they don’t actually have a good r… | (Also, Tennessee’s record keeping and their two hundred years of using biological LLM confabulation means they don’t actually have a g… | ↗Not briefed |
arxiv arxiv-ai | 1mo | When Should an In-Context Learner Expand Its Hypothesis Space? | Learning systems adapt quickly inside a familiar family of models. The harder step comes earlier: deciding, from observations that cou… | ↗Not briefed |
hn hn-ai | 1mo | GoatedContext: Cross-repo, local preference context and cross-agent injection | — | ↗Not briefed |
bluesky bluesky-ai | 1mo | Knowledge. But not wisdom. openai.com/index/sharin... | Knowledge. But not wisdom. openai.com/index/sharin... | ↗Not briefed |
arxiv arxiv-ai | 1mo | Secure-CUA: Controlling Untrusted Influence in Computer-Use Agents | Computer-use agents (CUAs) perform tasks across applications (such as desktops, mobile apps, and web browsers) by observing graphical… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Boundary-Free Contextual Biasing: Depth-Adaptive Gating and Reading-Space Matching for Unsegmented Languages | Contextual biasing supplies an ASR system with a list of expected words at inference time, but existing methods rely on word boundarie… | ↗Not briefed |
arxiv arxiv-ai | 1mo | LLM-Enabled UAV Dispatch: A System-Level Survey and Taxonomy | Unmanned aerial vehicle (UAV) dispatch is beginning to move beyond isolated path planning and optimization-driven resource allocation… | ↗Not briefed |
arxiv arxiv-ai | 1mo | BanglaRhet: Benchmarking Classical and Transformer Models for Rhetorical and Persuasion Detection in Bangla Political Speech | Political discourse often uses rhetorical and persuasive language to frame narratives, influence public opinion, and mobilize audience… | ↗Not briefed |
hn hn-ai | 1mo | Experimental MSU-1 streaming and standard-memory profiles | — | ↗Not briefed |
hn hn-ai | 1mo | Sharded, encrypted storage between friends over Yggdrasil | — | ↗Not briefed |
rss techmeme | 1mo | Filing: China-based Transsion, which commands 53% of Africa's mobile phone market by unit sales, is seeking to raise as much as ~$428M in a Hong Kong share sale (Sangmi Cha/Bloomberg) | Sangmi Cha / Bloomberg: Filing: China-based Transsion, which commands 53% of Africa's mobile phone market by unit sales, is seeking to… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Something poetic about the European model being the best at legal. There's a joke to be made here but we have to run it through the board t… | Something poetic about the European model being the best at legal. There's a joke to be made here but we have to run it through the bo… | ↗Not briefed |
bluesky bluesky-ai | 1mo | look, i'm not against openai releasing their proofs, but this might have been, like, the worst way possible to do it | look, i'm not against openai releasing their proofs, but this might have been, like, the worst way possible to do it | ↗Not briefed |
hn hn-ai | 1mo | Show HN: Rgpu – a PyTorch device whose tensors live on a remote GPU | rGPU runs GPU work on a remote NVIDIA machine while the application stays on the client. | ↗Not briefed |
rss techmeme | 1mo | Sources: the European Commission considers taxing big US tech companies through a broad levy on large corporations to avoid singling out individual companies (Financial Times) | Financial Times: Sources: the European Commission considers taxing big US tech companies through a broad levy on large corporations to… | ↗Not briefed |
arxiv arxiv-ai | 1mo | An Invariant Tangent-Angle Descriptor and a Band U-Net for 2D Fragment Adjacency Prediction | This paper addresses the prediction of adjacency between pairs of 2D fragments based on their contours. We improved the two-stage arch… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Right Number, Wrong State? Measuring Cross-Jurisdiction Substitution in LLM Recall of State Policy🔧 GPT-5 | When an LLM answers a state-specific policy question wrongly, it may be hallucinating, or it may be returning a real value that holds… | ↗Not briefed |
arxiv arxiv-ai | 1mo | DSReg: Provably Recovering Individual World Latents without Reconstruction | Methods that recover individual latent variables of the world, from nonlinear ICA to dictionary learning and causal representation lea… | ↗Not briefed |
arxiv arxiv-ai | 1mo | GeoPrior-Mamba: Structured Process Priors with Mamba for Fine-Resolution XCO2 Reconstruction | Reconstructing fine-resolution column-averaged dry-air CO2 (XCO2) fields from sparse satellite observations requires models to infer s… | ↗Not briefed |
hn hn-ai | 1mo | Show HN: Pareto Partisan? Relative Gains and Support for Public Policy | — | ↗Not briefed |
bluesky bluesky-ai | 1mo | Barn Find: Military grade FuckBaby, 80 years old, still boots up. A very efficient model. The soldiers called this model the Kid Sister. #… | Barn Find: Military grade FuckBaby, 80 years old, still boots up. A very efficient model. The soldiers called this model the Kid Sist… | ↗Not briefed |
bluesky bluesky-ai | 1mo | If you carry out all the matrix multiplications and whatnot to do the LLM torture experiments using nothing but a Turing machine made of th… | If you carry out all the matrix multiplications and whatnot to do the LLM torture experiments using nothing but a Turing machine made… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Iris-3B: Going Beyond the Latent with Pixel-Space Diffusion Training, Conversion and Fine-Tuning🔧 FLUX | Pixel-space diffusion models avoid the lossy VAE of latent models, which suggests an advantage on downstream tasks where fine-grained… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Das KI-Unternehmen hat ein internes KI-Modell auf tausende Probleme aus der Mathematik angesetzt. Für ein Zehntel davon hat es relevante Lö… | Das KI-Unternehmen hat ein internes KI-Modell auf tausende Probleme aus der Mathematik angesetzt. Für ein Zehntel davon hat es relevan… | ↗Not briefed |
bluesky bluesky-ai | 1mo | openai should release future math results as embeddings | openai should release future math results as embeddings | ↗Not briefed |
arxiv arxiv-ai | 1mo | Arctic Questions, Missing Answers: A Dataset and Benchmark for LLM Abstention in Arctic Science🔧 ChatGPT | Large language models (LLMs) should abstain from scientific multiple-choice questions when no option is valid, but frequent abstention… | ↗Not briefed |
rss latent-space | 1mo | [AINews] Quasi-Riemann-Hypothesis: OpenAI publishes 722 math papers solving 90 of the top 500 open math problems; “the most significant moment” in >100 years of mathematics | Our head hurts. | ↗Not briefed |
hn hn-ai | 1mo | Show HN: KORA Doctor – find LLM calls your agent may not have needed | — | ↗Not briefed |
bluesky bluesky-ai | 1mo | So what do you do if you're a lab that has a model that can solve problems at that level? For a start you pump the gas on it a bit, to see … | So what do you do if you're a lab that has a model that can solve problems at that level? For a start you pump the gas on it a bit, to… | ↗Not briefed |
bluesky bluesky-ai | 1mo | OpenAI releases new cookbook “To Serve Man” | OpenAI releases new cookbook “To Serve Man” | ↗Not briefed |