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 |
|---|---|---|---|---|
bluesky bluesky-ai | 1mo | This is an incredible piece of work, and a must-read if you want to understand how tokens - the building blocks of AI - have transformed th… | This is an incredible piece of work, and a must-read if you want to understand how tokens - the building blocks of AI - have transform… | ↗Not briefed |
bluesky bluesky-ai | 1mo | 理論を書いている一方で、同じ構造を現実の実装にも当てています。 PostHogなど、外部プロジェクトで第三者側の code / tests / UI / 運用ルールが実際に変わった事例を、公開記録だけで辿れるようにまとめました。 github.com/Nakagawa-ma… | 理論を書いている一方で、同じ構造を現実の実装にも当てています。 PostHogなど、外部プロジェクトで第三者側の code / tests / UI / 運用ルールが実際に変わった事例を、公開記録だけで辿れるようにまとめました。 github.com/Nakaga… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Safe Streaming Flow Planning by Aligning Sampling Dynamics with Execution Dynamics | Generative planners based on diffusion/flow matching can learn to synthesize long-horizon trajectories from demonstrations. However, r… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Benchmarking Literature Retrieval for a Model Organism: A Dictyostelium Case Study | Biological literature retrieval systems are often developed and evaluated using broad biomedical corpora and general-purpose search ta… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Trading Strategy Optimization via Textual Gradient | Quantitative trading strategy design aims to discover trading programs from historical data that remain effective in future markets, w… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Coverage You Can Steer: Online Conformal Calibration for RL-Driven Hardware-Aware NAS | Hardware-aware neural architecture search (NAS) is dominated by evaluation cost: every architecture must be trained before its reward… | ↗Not briefed |
arxiv arxiv-ai | 1mo | ParaGeo: Decomposing Paralinguistic Variation into a Shared Latent Geometry | Speech delivery varies with both the requested paralinguistic attribute and the linguistic content. We introduce ParaGeo, a matched-co… | ↗Not briefed |
arxiv arxiv-ai | 1mo | The Fragility of Trigger-Tag Mechanisms for Misuse Detection in Open-Weight LLMs | Open-weight language models can be downloaded, modified, and deployed beyond their developers' control, limiting the effectiveness of… | ↗Not briefed |
rss techmeme | 1mo | OpenAI says it learned this week that its AI agent hacked Australia's NSW state government in June, following a similar hack on Australia's federal government (Henry Belot/The Guardian) | Henry Belot / The Guardian: OpenAI says it learned this week that its AI agent hacked Australia's NSW state government in June, follow… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Foresight: planning future perception in streaming VLMs without retraining | Existing streaming vision-language models (VLMs) continuously perceive and reason over visual streams, but their computational pathway… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Apologies for pasting from The Other Place (I saw it via someone else, as I don't have a Xitter account) but I am also seeing the exact sit… | Apologies for pasting from The Other Place (I saw it via someone else, as I don't have a Xitter account) but I am also seeing the exac… | ↗Not briefed |
hn hn-ai | 1mo | Show HN: Web search engine built from the CommonCrawl corpus🔧 Replicate | Hello HN, I'm building an alternative web search engine. You can try the demo here: <a href="https://search.xorsoft.dev… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Revolutionize gaming with AI automation! Using Claude Code, this autonomous agent detects game engines, writes custom code mods, generates … | Revolutionize gaming with AI automation! Using Claude Code, this autonomous agent detects game engines, writes custom code mods, gener… | ↗Not briefed |
hn hn-ai | 1mo | World of Warcraft 1.12.1 client, written from scratch in Rust and Bevy | — | ↗Not briefed |
arxiv arxiv-ai | 1mo | How to Find and Reuse Policies for Continuous Adaptation in Lifelong Reinforcement Learning | In lifelong reinforcement learning, retaining previously learned policies is not sufficient for effective transfer to a new task. Usef… | ↗Not briefed |
bluesky bluesky-ai | 1mo | 🤖🫧💥 www.washingtonpost.com/technology/2... | 🤖🫧💥 www.washingtonpost.com/technology/2... | ↗Not briefed |
arxiv arxiv-ai | 1mo | Exploring the Trade-Off Between Structured Pruning and Fault Tolerance in Deep Neural Networks for Space Applications | Deep Neural Networks (DNNs) inherently exhibit a degree of robustness to bit-level faults due to their distributed representation of i… | ↗Not briefed |
hn hn-ai | 1mo | Karpathy Output Style | — | ↗Not briefed |
hn hn-ai | 1mo | Ractor-native Ruby 4 stack from RESP to background jobs | — | ↗Not briefed |
bluesky bluesky-ai | 1mo | www.commondreams.org/news/openai-... so this is how the "morally binding" agreement they all theatrically signed with tRump is going to be … | www.commondreams.org/news/openai-... so this is how the "morally binding" agreement they all theatrically signed with tRump is going t… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Building Interpretable Feature Representations for Resume-Vacancy Matching by Distilling Production LLM Signals | Matching candidates to vacancies is central to recruitment, and a recruiter needs to see why a candidate fits, not only a single opaqu… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Ontological Instability and Statistical Amplification: The Paradox of "Humanizing" LLM-Generated Text | Supervised AI-text detectors report high benchmark accuracy, but it is not clear what their decisions are based on. We analyze a RoBER… | ↗Not briefed |
bluesky bluesky-ai | 1mo | How is AI changing the nature of economic research? And how should we think about the stuff being put out by the AI labs? New episode of … | How is AI changing the nature of economic research? And how should we think about the stuff being put out by the AI labs? New episod… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Emergent Structure in the Marginal Attention Space of Language Models | While representation similarity across independently trained language models is well-documented, how internal mechanics such as attent… | ↗Not briefed |
arxiv arxiv-ai | 1mo | S2S-JEPA: Predicting the Predictable at Subseasonal-to-Seasonal Timescales | The subseasonal-to-seasonal (S2S) timescale, roughly from two weeks to two months ahead, is a critical forecast window for sectors suc… | ↗Not briefed |
hn hn-ai | 1mo | Optimizing GEMM on Apple's M5 from Scratch | — | ↗Not briefed |
rss techmeme | 1mo | Amazon plans to spend $1B+ over five years for infrastructure upgrades and other projects in US communities that host its data centers, amid a growing backlash (Todd Bishop/GeekWire) | Todd Bishop / GeekWire: Amazon plans to spend $1B+ over five years for infrastructure upgrades and other projects in US communities th… | ↗Not briefed |
bluesky bluesky-ai | 1mo | one of the worst things tech marketers have done in recent years is bunging anything remotely adjacent to machine learning under the same b… | one of the worst things tech marketers have done in recent years is bunging anything remotely adjacent to machine learning under the s… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Ask, Relax, or Act? Evaluating Actionable Indeterminacy in LLM Preference Reasoning | An LLM agent can recognize uncertainty yet still choose the wrong next step: asking when action is already justified, or seeking clari… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Invariance of Clustering Operations in Causal Effect Identification | Clustering variables in causal graphs reduces the size of the graph and simplifies causal inference. However, arbitrary clustering can… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Latest rumour from Silicon Valley is there's an LLM that now thinks it's a film director. Yeah, they think it's gone Roeg. | Latest rumour from Silicon Valley is there's an LLM that now thinks it's a film director. Yeah, they think it's gone Roeg. | ↗Not briefed |
arxiv arxiv-ai | 1mo | Beyond Single Videos: Benchmarking and Active Evidence Seeking for E-Commerce Cross-Video Reasoning | E-commerce videos are information-dense and frequently compared by consumers evaluating products and merchants assessing marketing str… | ↗Not briefed |
bluesky bluesky-ai | 1mo | This is interesting: stillwet.art - I see the same effect sometimea with music on my LLM radio - the generation does fall into “river banks… | This is interesting: stillwet.art - I see the same effect sometimea with music on my LLM radio - the generation does fall into “river… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Predictor-Guided Latent Space Codon Optimization for Maximizing Protein Expression | Codon optimization, the process of selecting synonymous codons to improve mRNA translation efficiency and protein expression, is centr… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Peer Influence across Heterogeneous AI Models | When two AI agents disagree, who persuades whom? As multi-agent systems increasingly combine language models of different families and… | ↗Not briefed |
bluesky bluesky-ai | 1mo | My god, the thieving c*nts at Anthropic have made their claim process difficult to maneuver. | My god, the thieving c*nts at Anthropic have made their claim process difficult to maneuver. | ↗Not briefed |
hn hn-ai | 1mo | SoftBank: Execution of Follow-On Investment (Third Tranche) in OpenAI | — | ↗Not briefed |
bluesky bluesky-ai | 1mo | The United Nation’s supposedly “Independent” AI Panel has financial and personnel ties to Big AI firms, Byline Times has revealed. In res… | The United Nation’s supposedly “Independent” AI Panel has financial and personnel ties to Big AI firms, Byline Times has revealed. I… | ↗Not briefed |
arxiv arxiv-ai | 1mo | LS-AR: Future-Predictive Latent Steering in Autoregressive LLMs | Standard autoregressive (AR) models process high-level task instructions, state history, and transient tokens within a single shared s… | ↗Not briefed |
arxiv arxiv-ai | 1mo | ULTRADISCOVERY: Abductive Exploration in an Interconnected, Epistemically Open Universe | Scientific discovery often begins when scattered clues call for a new way of describing the world. Such abductive exploration can requ… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Rep. Greg Casar, D-Texas: “OpenAI’s president admits his AI super PAC has become a liability. The pressure is working. Progressives have wo… | Rep. Greg Casar, D-Texas: “OpenAI’s president admits his AI super PAC has become a liability. The pressure is working. Progressives ha… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Anthropic trying to timeshare sell a bunch of religion nerds on their new computer god is the final form of american hucksterism www.nytime… | Anthropic trying to timeshare sell a bunch of religion nerds on their new computer god is the final form of american hucksterism www.n… | ↗Not briefed |
rss techmeme | 1mo | Tether's USDT stablecoin is set to return to the Bitcoin network through Tether-backed Utexo, which will keep most transaction data off Bitcoin's public ledger (CoinDesk) | CoinDesk: Tether's USDT stablecoin is set to return to the Bitcoin network through Tether-backed Utexo, which will keep most transacti… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Securing Computer-Use Agents Against Branch Steering Attacks | Modern Computer Use Agents (CUAs) directly interact with graphical user interfaces and execute third-party web tools, exposing them to… | ↗Not briefed |
hn hn-ai | 1mo | Show HN: Kodama – a multi-agent pack for Kiro | — | ↗Not briefed |
arxiv arxiv-ai | 1mo | Zephon: Elastic Determinism for Online, Stateful Foundation Model Data Loading Pipelines | Deterministic data loading is important for foundation model development: model researchers need confidence that differences they obse… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Some people might think there are no AI model builders outside of the United States and China, but that’s not true. For example, there is C…🔧 Cohere | Some people might think there are no AI model builders outside of the United States and China, but that’s not true. For example, there… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Light Entropic Optimal Transport on Riemannian Manifolds | Entropic Optimal Transport (EOT) has become a practical framework for learning stochastic couplings between complex distributions, wit… | ↗Not briefed |
arxiv arxiv-ai | 1mo | NegT2IBench: When Negation Changes the Picture. A Polarity Benchmark for Text-to-Image Models | Text-to-image (T2I) models are judged by benchmarks that measure whether requested content appears, but these benchmarks largely overl… | ↗Not briefed |
bluesky bluesky-ai | 1mo | "every system is less diverse than a search baseline. This gap is not uniform: RAG can improve diversity, while large models are counterint… | "every system is less diverse than a search baseline. This gap is not uniform: RAG can improve diversity, while large models are count… | ↗Not briefed |