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 | 29d | On Global Joint Lower Bounds in Zeroth-Order Nonconvex Optimization | Recent work has sought to transfer first-order nonconvex optimization guarantees to zeroth-order methods through gradient estimators,… | ↗Not briefed |
hn hn-ai | 29d | AOSP Source Releases Are So Late Developers Disassemble Libraries to Report Bugs | — | ↗Not briefed |
arxiv arxiv-ai | 29d | AWARE-RG: Abstraction Aware Explanation of Resiliency Graphs for Cyber Physical Systems | Understanding the security posture of cyber physical systems (CPS) requires reasoning across both information technology (IT) and oper… | ↗Not briefed |
bluesky bluesky-ai | 29d | "we're seeing signs of life for AI worms" = assume the position while we hack your bank account and your fridge, suckerssss How is this ok… | "we're seeing signs of life for AI worms" = assume the position while we hack your bank account and your fridge, suckerssss How is th… | ↗Not briefed |
hn hn-ai | 29d | AI made feasibility almost trivial. Not touched: value, viability, or usability | — | ↗Not briefed |
bluesky bluesky-ai | 29d | Model: Alma Experimentals PDXL sr1.1 Civitai: civitai.red/user/AlmaScar Ko-Fi: ko-fi.com/almascar #beach #redhead #bikini #popsicle #beaut…🔧 Civitai | Model: Alma Experimentals PDXL sr1.1 Civitai: civitai.red/user/AlmaScar Ko-Fi: ko-fi.com/almascar #beach #redhead #bikini #popsicle #… | ↗Not briefed |
arxiv arxiv-ai | 29d | Selective Reasoning for Typed Decision Models: When to Start, Read, and Stop | Typed decision models need an interface that can use additional reasoning while retaining candidate outputs. We train a shared candida… | ↗Not briefed |
bluesky bluesky-ai | 29d | Anthropic glazers are in a league of their own tbh | Anthropic glazers are in a league of their own tbh | ↗Not briefed |
bluesky bluesky-ai | 29d | The AI industry has spent years scraping everything we've ever written, said, or posted without asking. A TV in OpenAI's office gets bugged… | The AI industry has spent years scraping everything we've ever written, said, or posted without asking. A TV in OpenAI's office gets b… | ↗Not briefed |
arxiv arxiv-ai | 29d | Model What Matters: Decision-Guided Selective World Modeling for Language Agents | Reinforcement learning provides a direct mechanism for optimizing task objectives in large language model-driven agents, but task rewa… | ↗Not briefed |
arxiv arxiv-ai | 29d | Synthetic reaction data from transformation rules for graph-to-sequence pretraining in biodegradation pathway prediction | The scarcity of observed biodegradation reactions constrains the training of product and pathway predictors. We present enviGraph, a f… | ↗Not briefed |
hn hn-ai | 29d | Show HN: Agent-trials – a smoke-test runner for coding agent CLIs | — | ↗Not briefed |
rss techmeme | 29d | A look at a coordinated theft of ~$100M in crypto using hardware backdoors planted in Ledger wallets sold by Malaysia-based authorized reseller CryptoBilis (Kim Zetter/ZERO DAY) | Kim Zetter / ZERO DAY: A look at a coordinated theft of ~$100M in crypto using hardware backdoors planted in Ledger wallets sold by Ma… | ↗Not briefed |
bluesky bluesky-ai | 29d | Yeah it's definitely just anthropic sure. Just $5bn more dollars bro I promise I only need $7bn more dollars and we'll be profitable I swea… | Yeah it's definitely just anthropic sure. Just $5bn more dollars bro I promise I only need $7bn more dollars and we'll be profitable I… | ↗Not briefed |
bluesky bluesky-ai | 29d | Just saw some AI booster cope like "it's just anthropic that aren't profitable not the whole sector" and ooh boy is that schadenfreude goin… | Just saw some AI booster cope like "it's just anthropic that aren't profitable not the whole sector" and ooh boy is that schadenfreude… | ↗Not briefed |
arxiv arxiv-ai | 29d | J-tau: A Fully Localized Japanese Benchmark for Tool-Agent-User Interaction | We present J-tau, the first publicly available Japanese benchmark for evaluating tool-using LLM agents in interactive customer-support… | ↗Not briefed |
bluesky bluesky-ai | 29d | Now you will have to start paying a premium for retro non internet connected ovens, toothbrushes and fridges www.ft.com/content/64aa... | Now you will have to start paying a premium for retro non internet connected ovens, toothbrushes and fridges www.ft.com/content/64aa... | ↗Not briefed |
arxiv arxiv-ai | 29d | Sparse Autoencoders Are a Shared Coordinate System for Symbolic Matching | Large language models (LLMs) hold broad knowledge in one set of weights, but lack interpretability and trustworthy reasoning, especial… | ↗Not briefed |
arxiv arxiv-ai | 29d | Eye, Robot: Learning Expressive Gaze Dynamics via a VLM-Guided Transformer-VAE | Gaze coordinates attention in human interaction through coupled eye, head, and eyelid motion. A humanoid robot must choose where to lo… | ↗Not briefed |
bluesky bluesky-ai | 29d | その上、メタの盗撮メガネで隠し撮りなんかされたら終わりです。一生気付かないまま、凌辱され続けるのです。卒業アルバムを裸にされた子たちは一生苦しむ。二次被害、三次被害を生み続ける。リベンジポルノの大量増殖版です。 犯人はただの遊びで軽い気持ちでやったことは、罪を償ったとしても、… | その上、メタの盗撮メガネで隠し撮りなんかされたら終わりです。一生気付かないまま、凌辱され続けるのです。卒業アルバムを裸にされた子たちは一生苦しむ。二次被害、三次被害を生み続ける。リベンジポルノの大量増殖版です。 犯人はただの遊びで軽い気持ちでやったことは、罪を償った… | ↗Not briefed |
bluesky bluesky-ai | 29d | Gemini 4 Argon debuts with a catch 🤑 Simon Willison's `llm-gemini` tool now supports Gemini 4 Argon, a new model with strong performance … | Gemini 4 Argon debuts with a catch 🤑 Simon Willison's `llm-gemini` tool now supports Gemini 4 Argon, a new model with strong perform… | ↗Not briefed |
rss techmeme | 29d | ASML says it cannot fully meet AI-driven demand for its chipmaking tools in 2027 and 2028, and is seeing "strong interest" in orders for 2029 and 2030 (Reuters) | Reuters: ASML says it cannot fully meet AI-driven demand for its chipmaking tools in 2027 and 2028, and is seeing “strong interest” in… | ↗Not briefed |
arxiv arxiv-ai | 29d | Risk-Controlled DFT Selection for Machine-Learned Interatomic Potentials | Automated materials discovery requires evaluating large pools of candidate configurations, with density functional theory (DFT) as a m… | ↗Not briefed |
hn hn-ai | 29d | LLMs Are Turing-Complete | — | ↗Not briefed |
arxiv arxiv-ai | 29d | Self-Conditioning Enables Deep Sequential Reasoning in Flow Matching | Recent work has explored visual reasoning through image generation, where models solve visual-reasoning problems by directly generatin… | ↗Not briefed |
hn hn-ai | 29d | Anthropic conjures threat that doesn't exist yet to stoke fears of AI apocalypse | — | ↗Not briefed |
bluesky bluesky-ai | 29d | In 1988 you would end up in jail for that … www.theguardian.com/technology/2... Check en.wikipedia.org/wiki/Morris_... | In 1988 you would end up in jail for that … www.theguardian.com/technology/2... Check en.wikipedia.org/wiki/Morris_... | ↗Not briefed |
hn hn-ai | 29d | Floco: Using Nix to put NPM and Yarn in a coffin | — | ↗Not briefed |
arxiv arxiv-ai | 29d | DAGFormer: Global Topology Prediction for Transformer Cross-Layer Connections | We propose DAGFormer, a Transformer language model whose cross-layer connections are planned by a jointly trained topology predictor.… | ↗Not briefed |
rss bing-news-ai | 29d | Anthropic says small artificial intelligence models pose new self-replication risks | Artificial intelligence company Anthropic warns that small models could soon copy themselves across everyday household devices. Accord… | ↗Not briefed |
rss bing-news-ai | 29d | Anthropic says small artificial intelligence models pose new self-replication risks | Artificial intelligence company Anthropic warns that small models could soon copy themselves across everyday household devices. Accord… | ↗Not briefed |
rss bing-news-ai | 29d | Anthropic says small artificial intelligence models pose new self-replication risks | Artificial intelligence company Anthropic warns that small models could soon copy themselves across everyday household devices. Accord… | ↗Not briefed |
bluesky bluesky-ai | 29d | Update: Now OpenAI registers as a victim - and Microsoft too, as a 10+ percent owner | Update: Now OpenAI registers as a victim - and Microsoft too, as a 10+ percent owner | ↗Not briefed |
hn hn-ai | 29d | LeanLLM: An LLM SDK for applications written in Lean 4 | — | ↗Not briefed |
hn hn-ai | 29d | OpenAI employee arrested after recording device found in office | — | ↗Not briefed |
hn hn-ai | 29d | Houthis expose limits of Saudi Arabia's arsenal | — | ↗Not briefed |
arxiv arxiv-ai | 29d | Are AI Decision Models Biased? A Case Study of Jev | Do AI models designed for fast, structured decisions exhibit the cognitive and social biases observed in human judgment? We evaluate J… | ↗Not briefed |
hn hn-ai | 29d | GitDrive – private GitHub repos as an encrypted, iCloud-style Finder drive | — | ↗Not briefed |
rss latent-space | 29d | [AINews] AI Engineer NYC: Gemini 4 Argon, Vals AI Benchmarks | Gemini 4 Argon went GA today, but we already gave it title story 2 weeks ago and most benchmarks, like CursorBench, simply independent… | ↗Not briefed |
rss techmeme | 29d | ASML reports Q3 net sales of €11.66B, vs. €11.72B est., a €4.96B net profit, vs. €4.1B est., and raises its 2026 net sales forecast from €43B-€44B to €44B-€45B (Arjun Kharpal/CNBC) | Arjun Kharpal / CNBC: ASML reports Q3 net sales of €11.66B, vs. €11.72B est., a €4.96B net profit, vs. €4.1B est., and raises its 2026… | ↗Not briefed |
arxiv arxiv-ai | 29d | Instruction-Conditioned Motion Control for Robotic Arms: Language-to-Code Decomposition and Reinforcement Learning | Instruction-Conditioned Motion Control (ICMC) separates parameter specification, instruction generation, and code synthesis for langua… | ↗Not briefed |
bluesky bluesky-ai | 29d | OpenAI is pissed off that someone copied their words and ideas without their permission? | OpenAI is pissed off that someone copied their words and ideas without their permission? | ↗Not briefed |
arxiv arxiv-ai | 29d | Divergent Paths: Learning an Outcome-Predictive Representation of LLM Agent Trajectories | In this work, we reframe agentic trajectories as paths through a shared embedding space of actions and environmental observations. We… | ↗Not briefed |
arxiv arxiv-ai | 29d | Are We Really Measuring Forgetting? Pitfalls in Evaluating Machine Unlearning for Multimodal LLMs | As machine unlearning for multimodal large language models (MLLMs) rapidly advances, rigorous evaluation of its effectiveness becomes… | ↗Not briefed |
bluesky bluesky-ai | 29d | like if the fridge has so much RAM and storage and processing power that it can run an LLM, imagine what actual hackers (who do amazing thi… | like if the fridge has so much RAM and storage and processing power that it can run an LLM, imagine what actual hackers (who do amazin… | ↗Not briefed |
bluesky bluesky-ai | 29d | 本件、妙にバズってしまったので補足。 Anthropic社への怒りの声も観測されますが、本件に限っては筋近いではないかと思います。 ●Anthropicはオープンウェイトモデルを出していない。（つまりAnthropicは本件については当事者ではない） ●今回の記事は、AI… | 本件、妙にバズってしまったので補足。 Anthropic社への怒りの声も観測されますが、本件に限っては筋近いではないかと思います。 ●Anthropicはオープンウェイトモデルを出していない。（つまりAnthropicは本件については当事者ではない） ●今回の記… | ↗Not briefed |
bluesky bluesky-ai | 29d | like, if your exploit kit has an LLM in it, yes, all attack surfaces that are increasingly vulnerable to all exploit kits are more vulnerab… | like, if your exploit kit has an LLM in it, yes, all attack surfaces that are increasingly vulnerable to all exploit kits are more vul… | ↗Not briefed |
arxiv arxiv-ai | 29d | Twenty Days Later: A Pre-Registered Test of History-Based Forecasts of Brand Mentions in Gemini Answers | Counts of which brands a large language model names are useful only if they hold long enough to act on. We test this with a pre-regist… | ↗Not briefed |
arxiv arxiv-ai | 29d | A Graph-Based Analysis of Comparability, Scale, and Attribution in Private Transformer Inference | Rising privacy concerns around large language models have accelerated research in private Transformer inference (PTI). Synthesizing th… | ↗Not briefed |
bluesky bluesky-ai | 29d | obviously Anthropic is reckless and i agree with you about the framing. ...and the longtime cybersecurity worker in me is irritated that … | obviously Anthropic is reckless and i agree with you about the framing. ...and the longtime cybersecurity worker in me is irritated… | ↗Not briefed |